• Jean Francois Puget - Distinguished Engineer, Machine Learning and Optimization; Kaggle Grandmaster
• Horst Samulowitz - Research Manager and Research Staff Member in Artificial Intelligence, IBM Research
• Marius Danciu - Senior Technical Staff Member (STSM), Watson Machine Learning
• Greg Filla - Offering Manager, AutoAI, Watson Studio and Watson Machine Learning
• Yin Chen - Offering Manager, Deep Learning, Watson Studio and Watson Machine Learning
• Julianna Delua - Portfolio Manager, Data and AI

------------------------------
Christina Howell
------------------------------
#WatsonStudio

Hello, great news, my name is Emmanuel from Mexico and I want to make my career plan as a data scientist
In studio what recommend me for data visualization, I used Cognos Analytics but frequently the service was not available and the environment is very slow, I would like to connect the IBM Maximo tool with Watson and show information about incidents, assets and basically everything,

you very much

------------------------------
Emmanuel Gonzalez Sanchez
------------------------------

------Original Message------

Hello, great news, my name is Emmanuel from Mexico and I want to make my career plan as a data scientist
In studio what recommend me for data visualization, I used Cognos Analytics but frequently the service was not available and the environment is very slow, I would like to connect the IBM Maximo tool with Watson and show information about incidents, assets and basically everything,

you very much

------------------------------
Emmanuel Gonzalez Sanchez
------------------------------

#WatsonStudio

Hi Christina,

I would like to ask about the current differences between versions of Watson Studio (Cloud, Local & Desktop) and why these versions are different now. Do you have a plan in the near future to align all the 3 versions so that they all have the same ? 

Thanks a lot.
Ahmed Lashin.

------------------------------
Ahmed Lashin
------------------------------

Hi Ahmed!

We recently shipped Watson Studio Local 2.0 with exactly that goal, bring the cloud and local version together and aligned! You can learn more here: https://medium.com/ibm-watson/accelerate-your-data-science-life-cycle-using-ibm-watson-studio-2-0-808da0ca1d87

Desktop will always be a little bit different - and the goal here is to have a native experience that is great for Desktop, optimized to run on macOS and Windows.

------------------------------
ARMAND RUIZ GABERNET
------------------------------

Original Message:
Sent: Thu May 30, 2019 01:36 PM
From: Ahmed Lashin
Subject: AMA with Watson Studio Data Scientists on June 13

Hi Christina,

I would like to ask about the current differences between versions of Watson Studio (Cloud, Local & Desktop) and why these versions are different now. Do you have a plan in the near future to align all the 3 versions so that they all have the same ? 

Thanks a lot.
Ahmed Lashin.

------------------------------
Ahmed Lashin
------------------------------

@Armand Ruiz ​are WSL and WSC going to align in terms of UI (i.e., how one accesses projects, deployment, admin consoles)? I admit that I haven't had time to play with WSL 2.0 yet, but as of 1.2.x the user experience feels quite different to me.

------------------------------
CHERYLANNE HOWARD
------------------------------

Original Message:
Sent: Fri May 31, 2019 06:55 PM
From: ARMAND RUIZ GABERNET
Subject: AMA with Watson Studio Data Scientists on June 13

Hi Ahmed!

We recently shipped Watson Studio Local 2.0 with exactly that goal, bring the cloud and local version together and aligned! You can learn more here: https://medium.com/ibm-watson/accelerate-your-data-science-life-cycle-using-ibm-watson-studio-2-0-808da0ca1d87

Desktop will always be a little bit different - and the goal here is to have a native experience that is great for Desktop, optimized to run on macOS and Windows.

------------------------------
ARMAND RUIZ GABERNET
------------------------------


Original Message:
Sent: Thu May 30, 2019 01:36 PM
From: Ahmed Lashin
Subject: AMA with Watson Studio Data Scientists on June 13

Hi Christina,

I would like to ask about the current differences between versions of Watson Studio (Cloud, Local & Desktop) and why these versions are different now. Do you have a plan in the near future to align all the 3 versions so that they all have the same ? 

Thanks a lot.
Ahmed Lashin.

------------------------------
Ahmed Lashin

Hi @CHERY HOWARD - Yes!  2.0 was a major step forward towards this alignment.  We are working very hard to deliver a common experience across Watson Studio Cloud and Local. 

Thanks - ​

------------------------------
GREG FILLA
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:29 AM
From: CHERYLANNE HOWARD
Subject: AMA with Watson Studio Data Scientists on June 13

@Armand Ruiz ​are WSL and WSC going to align in terms of UI (i.e., how one accesses projects, deployment, admin consoles)? I admit that I haven't had time to play with WSL 2.0 yet, but as of 1.2.x the user experience feels quite different to me.

------------------------------
CHERYLANNE HOWARD
------------------------------


Original Message:
Sent: Fri May 31, 2019 06:55 PM
From: ARMAND RUIZ GABERNET
Subject: AMA with Watson Studio Data Scientists on June 13

Hi Ahmed!

We recently shipped Watson Studio Local 2.0 with exactly that goal, bring the cloud and local version together and aligned! You can learn more here: https://medium.com/ibm-watson/accelerate-your-data-science-life-cycle-using-ibm-watson-studio-2-0-808da0ca1d87

Desktop will always be a little bit different - and the goal here is to have a native experience that is great for Desktop, optimized to run on macOS and Windows.

------------------------------
ARMAND RUIZ GABERNET


Original Message:
Sent: Thu May 30, 2019 01:36 PM
From: Ahmed Lashin
Subject: AMA with Watson Studio Data Scientists on June 13

Hi Christina,

I would like to ask about the current differences between versions of Watson Studio (Cloud, Local & Desktop) and why these versions are different now. Do you have a plan in the near future to align all the 3 versions so that they all have the same ? 

Thanks a lot.
Ahmed Lashin.

------------------------------
Ahmed Lashin

Dear Christina,

I would like to ask you about the differences between the current versions of Watson Studio (Cloud, Local and Desktop).



------------------------------
Ahmed Lashin
------------------------------

Hope to know more about the current state and future direction of AutoAI/AutoML.

@Tian  You can find an overvi ​ew of auto ai here: https://www.ibm.com/cloud/watson-studio/autoai

It has been GAed end of May and you can now try it.

We plan to add more features and capabilities on a regular basis, and also improve performance.  Stay tuned.

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Original Message:
Sent: Fri May 31, 2019 04:08 AM
From: TIAN HUAT TAN
Subject: AMA with Watson Studio Data Scientists on June 13

Hope to know more about the current state and future direction of AutoAI/AutoML.

@TIAN HUAT TAN  we are moving to unify user experience, make it easier for data scientists to further finetune models from AutoAI​​ and integrate other AI assets from IBM.  We welcome more feedback here.  Check out https://www.youtube.com/watch?v=ILCsbh9IKT0

------------------------------
Julianna Delua
------------------------------

Original Message:
Sent: Fri May 31, 2019 04:08 AM
From: TIAN HUAT TAN
Subject: AMA with Watson Studio Data Scientists on June 13

Hope to know more about the current state and future direction of AutoAI/AutoML.

It is nice to meet you and to know what your background is. I know I can learn a lot from this group. I am a new System Administrator located at Columbia, Missouri Center. I am on a learning path to know more about what IBM offers and to understand more in-depth their services so that I can perform my job better. I will be able to guide a Client when they need help in a specific area.  

Thank You,
Yolanda Hernandez

------------------------------
Yolanda Hernandez
System Admin Apprentice
IBM
Columbia, MO.
573-219-5895
------------------------------

@Yolanda your help is most welcome. ​  Feel free to ping us here or via IBM internal means as you see fit.

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Original Message:
Sent: Fri May 31, 2019 09:02 AM
From: Yolanda Hernandez
Subject: AMA with Watson Studio Data Scientists on June 13

It is nice to meet you and to know what your background is. I know I can learn a lot from this group. I am a new System Administrator located at Columbia, Missouri Center. I am on a learning path to know more about what IBM offers and to understand more in-depth their services so that I can perform my job better. I will be able to guide a Client when they need help in a specific area.  

Thank You,
Yolanda Hernandez

------------------------------
Yolanda Hernandez
System Admin Apprentice
IBM
Columbia, MO.
573-219-5895
------------------------------

Hello Christina!
Is it possible to use Watson Assistant for database analysis in order to get more appropriate responses? If yes, how?
If not, can we create a communication channel between Watson Assistant and Watson Studio?  If yes, how?

Regards!

------------------------------
Konan Jean-Claude Kouassi
Practice Makes Perfect!
------------------------------

@Konan Jean-Claude Kouassi  We can create associated services from Watson Studio using API for custom apps and others. We need to learn a bit more about your use cases  but you can learn more here including Watson Discovery   https://dataplatform.cloud.ibm.com/docs/content/wsj/getting-started/assoc-services.html?linkInPage=true 

If you enabled Watson Assistant with API, we should be able to handle and should verify with our implementation.  I found this in github https://github.com/IBM/watson-nao-robot  to discuss integration of Watson Studio and Watson Assistant

Watson Discovery
With the Watson Discovery service, you can use APIs in notebooks to add a cognitive search and content analytics engine to applications to identify patterns, trends and actionable insights that drive better decision-making. Securely unify structured and unstructured data with pre-enriched content, and use a simplified query language to eliminate the need for manual filtering of results. See Getting started tutorial. ​

------------------------------
Julianna Delua
------------------------------

Original Message:
Sent: Wed June 05, 2019 06:02 PM
From: Konan Jean-Claude Kouassi
Subject: AMA with Watson Studio Data Scientists on June 13

Hello Christina!
Is it possible to use Watson Assistant for database analysis in order to get more appropriate responses? If yes, how?
If not, can we create a communication channel between Watson Assistant and Watson Studio?  If yes, how?

Regards!

------------------------------
Konan Jean-Claude Kouassi
Practice Makes Perfect!
------------------------------

Hi, 

Can you advise on how  to work collaboratively with WSC within a team ? For instance :
- each one his own NB ?
- work separately on different chunk ?
- how do you merge work ?
- how do you share intermediate outputs of your 'pipeline'...
...
What is your POV vs WSC ?

------------------------------
MICHEL LE
------------------------------

@Michel what is WSC ? ;)  Assuming it is Watson Studio, answer may depend on who you work with.  In my experience, best is to work alone on a given notebook.  When you are done you commit it and other team member can fork it and make their changes. You share output via assets in your project. ​

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Original Message:
Sent: Wed June 12, 2019 11:39 AM
From: MICHEL LE
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, 

Can you advise on how  to work collaboratively with WSC within a team ? For instance :
- each one his own NB ?
- work separately on different chunk ?
- how do you merge work ?
- how do you share intermediate outputs of your 'pipeline'...
...
What is your POV vs WSC ?

------------------------------
MICHEL LE
------------------------------

Hi, 

Watson Studio Cloud

------------------------------
MICHEL LE
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:20 AM
From: JEAN FRANCOIS PUGET
Subject: AMA with Watson Studio Data Scientists on June 13

@Michel what is WSC ? ;)  Assuming it is Watson Studio, answer may depend on who you work with.  In my experience, best is to work alone on a given notebook.  When you are done you commit it and other team member can fork it and make their changes. You share output via assets in your project. ​

------------------------------
JEAN FRANCOIS PUGET
------------------------------


Original Message:
Sent: Wed June 12, 2019 11:39 AM
From: MICHEL LE
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, 

Can you advise on how  to work collaboratively with WSC within a team ? For instance :
- each one his own NB ?
- work separately on different chunk ?
- how do you merge work ?
- how do you share intermediate outputs of your 'pipeline'...
...
What is your POV vs WSC ?

------------------------------
MICHEL LE

Hi Michel, In WSC every collaborator with Editor or Admin role can create their own Notebook in the same project, and each Notebook are assigned separate environment so the team can work concurrently, intermediate outputs can be saved in cloud object storage , database and many other data sources, which can be shared as project assets through supported connections: https://dataplatform.cloud.ibm.com/docs/content/wsj/manage-data/conn_types.html

Here is a video of how to add collaborator in WSC: https://www.youtube.com/watch?v=nEAlXBDjM08

------------------------------
YIN CHEN
------------------------------

Original Message:
Sent: Wed June 12, 2019 11:39 AM
From: MICHEL LE
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, 

Can you advise on how  to work collaboratively with WSC within a team ? For instance :
- each one his own NB ?
- work separately on different chunk ?
- how do you merge work ?
- how do you share intermediate outputs of your 'pipeline'...
...
What is your POV vs WSC ?

------------------------------
MICHEL LE
------------------------------

I am Jean-François Puget, currently technical lead for machine learning offerings, which includes Watson Studio, Watson Machine Learning, and Watson Openscale.  I also practice machine learning on customer projects and IBM projects.  Last, I enter machine learning competitions on Kaggle where I am a Competion Granmaster and also top Discussion Grandmaster.  My Kaggle handle is CPMP.  Machine Learning has been a lifetime topic of interest for me as I hold a PhD in that field from previous millennium.

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Hello, everyone.  I am Julianna Delua, Data and AI SME, responsible for the Watson Studio family.  Look forward to our discussion today.

------------------------------
Julianna Delua
------------------------------

Hi, I am Horst Samulowitz from IBM Research and I am working on automating AI. The main idea is to leverage AI itself to automate it. To that end we have employed techniques from Multi-arm bandits, Reinforcement Learning to Mathematical Optimization in order to tackle problems such as automated model selection and feature engineering.

------------------------------
HORST SAMULOWITZ
------------------------------

Hi Horst,

How does AutoAI compare to Watson Machine Learning in terms of the techniques/algorithms used for model selection.

- Zeydy

------------------------------
Zeydy Ortiz
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:05 AM
From: HORST SAMULOWITZ
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I am Horst Samulowitz from IBM Research and I am working on automating AI. The main idea is to leverage AI itself to automate it. To that end we have employed techniques from Multi-arm bandits, Reinforcement Learning to Mathematical Optimization in order to tackle problems such as automated model selection and feature engineering.

------------------------------
HORST SAMULOWITZ
------------------------------

Hi @Zeydy Ortiz Watson ML ModelBuilder is only for Spark ML, and the automated workflow in it is very simple.  It only performs  ​basic data cleaning, then run different models concurrently, and present them in their merit order.

Auto AI goes way beyond.  First it leverages Python ecosystem instead of being limited to Spark ML.  It therefore can explore many more algorithms.  Second, also explores advanced feature engineering, something model builder was not doing at all.

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:12 AM
From: Zeydy Ortiz
Subject: AMA with Watson Studio Data Scientists on June 13

Hi Horst,

How does AutoAI compare to Watson Machine Learning in terms of the techniques/algorithms used for model selection.

- Zeydy

------------------------------
Zeydy Ortiz
------------------------------


Original Message:
Sent: Thu June 13, 2019 11:05 AM
From: HORST SAMULOWITZ
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I am Horst Samulowitz from IBM Research and I am working on automating AI. The main idea is to leverage AI itself to automate it. To that end we have employed techniques from Multi-arm bandits, Reinforcement Learning to Mathematical Optimization in order to tackle problems such as automated model selection and feature engineering.

------------------------------
HORST SAMULOWITZ

What does AMA stand for?

------------------------------
Jean-Georges Perrin / IBM Champion x11
------------------------------

Found it - Ask Me Anything :)

------------------------------
Jean-Georges Perrin / IBM Champion x11
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:12 AM
From: JeanGeorges Perrin
Subject: AMA with Watson Studio Data Scientists on June 13

What does AMA stand for?

------------------------------
Jean-Georges Perrin / IBM Champion x11
------------------------------

Hey @Jean-Georges Perrin It means Ask Me Anything ​

------------------------------
Julianna Delua
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:12 AM
From: JeanGeorges Perrin
Subject: AMA with Watson Studio Data Scientists on June 13

What does AMA stand for?

------------------------------
Jean-Georges Perrin / IBM Champion x11
------------------------------

Ask Me Anithyng about....

------------------------------
Paulo Klein
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:12 AM
From: JeanGeorges Perrin
Subject: AMA with Watson Studio Data Scientists on June 13

What does AMA stand for?

------------------------------
Jean-Georges Perrin / IBM Champion x11
------------------------------

Hey everyone!

I am Greg Filla, Product Manager for AutoAI.  I get to work with our incredible research, engineering, and design teams to make AutoAI a feature integrated with Watson Studio and Watson Machine Learning.  These are offerings from IBM that help teams get more value from their data specifically around building AI/ML solutions.  I previously worked as a data scientist at IBM and would have loved to use these tools to make my job easier :) Ask me anything!

------------------------------
Greg Filla
------------------------------

------------------------------
Ayush Anand
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:14 AM
From: GREG FILLA
Subject: AMA with Watson Studio Data Scientists on June 13

Hey everyone!

I am Greg Filla, Product Manager for AutoAI.  I get to work with our incredible research, engineering, and design teams to make AutoAI a feature integrated with Watson Studio and Watson Machine Learning.  These are offerings from IBM that help teams get more value from their data specifically around building AI/ML solutions.  I previously worked as a data scientist at IBM and would have loved to use these tools to make my job easier :) Ask me anything!

------------------------------
Greg Filla
------------------------------

You can use most popular and best open source to create customer machine learning models with Watson Studio.  This includes Deep learning frameworks such as Keras TensorFlow and Pytorch.  With these you can build models to detect spam provided you have enough training data.

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:17 AM
From: Ayush Anand
Subject: AMA with Watson Studio Data Scientists on June 13

Can Watson AI Platform be used to create ML and DL Algorithms that have a way with categorizing and extracting word patterns across sample text to identify which looks Spam?

Actually, I am a student and am in the way of creating an ML Model (preferably on Watson) to integrate it with my Action on Google Assistant to fight and categorize Fake News.

------------------------------
Ayush Anand
------------------------------


Original Message:
Sent: Thu June 13, 2019 11:14 AM
From: GREG FILLA
Subject: AMA with Watson Studio Data Scientists on June 13

Hey everyone!

I am Greg Filla, Product Manager for AutoAI.  I get to work with our incredible research, engineering, and design teams to make AutoAI a feature integrated with Watson Studio and Watson Machine Learning.  These are offerings from IBM that help teams get more value from their data specifically around building AI/ML solutions.  I previously worked as a data scientist at IBM and would have loved to use these tools to make my job easier :) Ask me anything!

------------------------------
Greg Filla

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003
------------------------------

------------------------------
martin Anderson
MD Pulse medic
ESSEx
7440787899
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003
------------------------------

@Martin Anderson  It is true that depending on the level of specialization, organizations need to add more layers.  For medical specific implementation, while you can use some of the modular components and services such as visual recognition a ​nd natural language classification, it would make sense to have medical specific data science layer where sub component can be training for specific personnel like checklists as well as personnel availability, etc for insight and prediction.  Watson Studio has been used extensively for healthcare and medical communities and Decision optimization for Watson Studio  https://www.ibm.com/us-en/marketplace/decision-optimization-for-watson-studio/details  can be used to take the machine learning prediction and further optimize your resources - fleet, personnel, equipment for clinical needs.

------------------------------
Julianna Delua
------------------------------

Original Message:
Sent: Thu June 13, 2019 12:11 PM
From: martin Anderson
Subject: AMA with Watson Studio Data Scientists on June 13

I'm thinking that IBM Watson uses a triple layer approach.

An industry layer, custom, and a public layer. Along with object storage.

Question:

As we are Advance life support, I and most of the team are senior clinicians. Do we need a fourth layer?

A training, Medical Data Science layer. OR one for the clinical/medocal products that can be built.

 Preferred Accessibility Setting:

Bullet Pointed and Task Orientated

Martin Anderson

Bloomberg Professor,

Clinical Alchemist Facebook

Clinical Business Architect NHS

Dev'Ops Service Integration Technology Partner Amazon | Google | Facebook

Clin'Ops Partner

PNP CC, BscEc, DIpHe Nursing, AutNurPra , ANP, NP, NMP, ENP, TICS.

Graduate Certificate In Electrocardiography

Microsoft Clinical Modern Business Partner

Microsoft Azure Developer Benefit Program

IBM Business Blockchain Badge Holder

020 8396 4356 07440787899 ONC

IBM Cloud X2 Badge Holder

IBM Data Science Practitioner

Pulse Medic, London South Bank University, The Technopark, SE1 6LN

------------------------------
martin Anderson
MD Pulse medic
ESSEx
7440787899
------------------------------


Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003

@Thorsen Gressling let me try to answer your questions.

(1) We support CSV files in Watson Studio, be it on cloud or local.  With the local version you can use the underlying filesystem as you see fit.​​ For cloud, is you question about the ability to create directories to save csv files?  We provide various ways to group and view your assets, that may be more convenient than a pseudo file system on the cloud.  You mention anaconda.  We provide runtimes with anaconda in Watson Studio.  Our notebooks are Jupyter notebooks.  And yes, you can access csv files from Jupyter notebooks in Watson Studio.  pandas.read_csv() is probably the most used function in Watson Studio notebooks ;)

(2) You can also use conda install, and if need be you can create your own Docker container with any additional software you need.

(3) Can you provide examples of Eclipse Plug Ins you find interesting?  Our take is that we separate the world of data science where one builds model, and the developer world where one consumes models.  The former can be done with Watson Studio, using Python ecosystem, or R, or SPSS, and the latter with Watson Machine Learning.  Once trained, models and pipelines are exported from Watson Studio to Watson ML.  There they can be deployed as web services, or run as batch against datasets.  They can also be applied to streamed data.  When deployed as a rest service, any kind of application and service can call the rest endpoint, and you can work with Java, or any other language you'd like.  There is no need to have data scientists work with Java nor to have developers work with Python ML/DL packages.

------------------------------
JEAN FRANCOIS PUGET
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003
------------------------------

Hi @Thorsten Gressling - thanks for your questions!

1. We support CSVs today for Watson Studio on Cloud (and with AutoAI).  The primary storage service is Cloud Object Storage (COS), which is a standard service across public cloud use cases no matter what cloud being used.  There are a number of benefits of using COS over a traditional filesystem, for example API access for any asset stored in a COS bucket.  We know this is a new skill set for data scientists moving from on prem to cloud so we offer libraries to help, e.g. Project-lib - https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/project-lib-python.html?audience=wdp&context=wdp

2.  This is not yet supported for Watson Studio Cloud, but we are exploring how to enable this in a secure way.  We do enable pip /conda installs at the environment customization level.  We also allow greater environment customization in Watson Studio Local which is a private cloud offering -as Jean Francois mentioned in his comment re: docker customization.

3. Do you mean open source / community built extensions for SPSS?  We do support this for SPSS Modeler outside of Watson Studio, we are planning to support similar open source integrations for Modeler flows in Watson Studio  - stay tuned for these updates.

------------------------------
GREG FILLA
------------------------------

Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003
------------------------------

This is a great news that we can ask questions to experts.

I too like to build my career in data science and I have started preparing for it but it is not much practical, So my question is how do I use watson studio for practicals on machine learning ?

What are the steps that to load the data set and use it for practice.

Thank you.

------------------------------
Priyanka Dhanraj Shinde
------------------------------

Original Message:
Sent: Thu June 13, 2019 12:35 PM
From: GREG FILLA
Subject: AMA with Watson Studio Data Scientists on June 13

Hi @Thorsten Gressling - thanks for your questions!

1. We support CSVs today for Watson Studio on Cloud (and with AutoAI).  The primary storage service is Cloud Object Storage (COS), which is a standard service across public cloud use cases no matter what cloud being used.  There are a number of benefits of using COS over a traditional filesystem, for example API access for any asset stored in a COS bucket.  We know this is a new skill set for data scientists moving from on prem to cloud so we offer libraries to help, e.g. Project-lib - https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/project-lib-python.html?audience=wdp&context=wdp

2.  This is not yet supported, but we are exploring how to enable this in a secure way.  We do enable pip /conda installs at the environment customization level.  We also allow greater environment customization in Watson Studio Local which is a private cloud offering.

3. Do you mean open source / community built extensions for SPSS?  We do support this for SPSS Modeler outside of Watson Studio, we are planning to support similar open source integrations for Modeler flows in Watson Studio  - stay tuned for these updates.

------------------------------
GREG FILLA
------------------------------


Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003

@Priyanka Dhanraj Shinde Many people use watson studio to get started and do more practicing.  We have many tutorials and this Data Refinery for shaping and visualizing data may be interesting  for you https://dataplatform.cloud.ibm.com/docs/content/wsj/refinery/dr_tutorial.html ​

You can get started by signing up https://dataplatform.cloud.ibm.com/registration/stepone?context=wdp&apps=all if you have not done so.  Good luck in building your data science career!


------------------------------
Julianna Delua
------------------------------

Original Message:
Sent: Thu June 13, 2019 12:43 PM
From: Priyanka Dhanraj Shinde
Subject: AMA with Watson Studio Data Scientists on June 13

This is a great news that we can ask questions to experts.

I too like to build my career in data science and I have started preparing for it but it is not much practical, So my question is how do I use watson studio for practicals on machine learning ?

What are the steps that to load the data set and use it for practice.

Thank you.

------------------------------
Priyanka Dhanraj Shinde
------------------------------


Original Message:
Sent: Thu June 13, 2019 12:35 PM
From: GREG FILLA
Subject: AMA with Watson Studio Data Scientists on June 13

Hi @Thorsten Gressling - thanks for your questions!

1. We support CSVs today for Watson Studio on Cloud (and with AutoAI).  The primary storage service is Cloud Object Storage (COS), which is a standard service across public cloud use cases no matter what cloud being used.  There are a number of benefits of using COS over a traditional filesystem, for example API access for any asset stored in a COS bucket.  We know this is a new skill set for data scientists moving from on prem to cloud so we offer libraries to help, e.g. Project-lib - https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/project-lib-python.html?audience=wdp&context=wdp

2.  This is not yet supported, but we are exploring how to enable this in a secure way.  We do enable pip /conda installs at the environment customization level.  We also allow greater environment customization in Watson Studio Local which is a private cloud offering.

3. Do you mean open source / community built extensions for SPSS?  We do support this for SPSS Modeler outside of Watson Studio, we are planning to support similar open source integrations for Modeler flows in Watson Studio  - stay tuned for these updates.

------------------------------
GREG FILLA


Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003

Thanks @JEAN FRANCOIS PUGET, @YIN CHEN, @Julianna Delua, @HORST SAMULOWITZ, and @Greg Filla for your responses and great discussion. 

A few general questions floated around that I gathered up below. Look forward to hearing your answers!

1. Which path of practical AI is better - go to university and get a PhD or learn on the job?
2. What's a day in your life look like? 
3. What hobbies do you have outside of datascience? 
4. What's it like working in IBM Research? 
5. What are your favorite places to learn about data science? 
6. Who are your mentors or 'famous' figures in industry? 

​​​​​​​​

------------------------------
Christina Howell
------------------------------

Original Message:
Sent: Thu June 13, 2019 12:43 PM
From: Priyanka Dhanraj Shinde
Subject: AMA with Watson Studio Data Scientists on June 13

This is a great news that we can ask questions to experts.

I too like to build my career in data science and I have started preparing for it but it is not much practical, So my question is how do I use watson studio for practicals on machine learning ?

What are the steps that to load the data set and use it for practice.

Thank you.

------------------------------
Priyanka Dhanraj Shinde
------------------------------


Original Message:
Sent: Thu June 13, 2019 12:35 PM
From: GREG FILLA
Subject: AMA with Watson Studio Data Scientists on June 13

Hi @Thorsten Gressling - thanks for your questions!

1. We support CSVs today for Watson Studio on Cloud (and with AutoAI).  The primary storage service is Cloud Object Storage (COS), which is a standard service across public cloud use cases no matter what cloud being used.  There are a number of benefits of using COS over a traditional filesystem, for example API access for any asset stored in a COS bucket.  We know this is a new skill set for data scientists moving from on prem to cloud so we offer libraries to help, e.g. Project-lib - https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/project-lib-python.html?audience=wdp&context=wdp

2.  This is not yet supported, but we are exploring how to enable this in a secure way.  We do enable pip /conda installs at the environment customization level.  We also allow greater environment customization in Watson Studio Local which is a private cloud offering.

3. Do you mean open source / community built extensions for SPSS?  We do support this for SPSS Modeler outside of Watson Studio, we are planning to support similar open source integrations for Modeler flows in Watson Studio  - stay tuned for these updates.

------------------------------
GREG FILLA


Original Message:
Sent: Thu June 13, 2019 11:52 AM
From: Thorsten Gressling
Subject: AMA with Watson Studio Data Scientists on June 13

Hi, I have three questions.

(1) CSV on file system. Still today the most relevant data ingestion is classic CSV. Fact, may we like it or not. One of the most asked features is a 'real file system' that I can work with. Will there be a classic FS that we can work with? Access from Jupyter? Showstopper: By now my students still use cloud9 with Jupyter and conda to work with a real OS on data science.

(2) Additional 'real' OS programs. We can use pip, but there are also more fundamental helper programs that must be installed in the OS, examples in our domain (Chemistry) are RDKit or Pymol. Is there a possibility to configure more down to the OS than just pip? 

(3) Using nodes/plugins from Eclipse-based Programs. One example is KNIME or RapidMinder where we have a lot of open source Java-based plugins supporting our business domain. Will SPSS be able to use and load these external plugins?

------------------------------
Thorsten Gressling
Business Leader Artificial Intelligence
ARS Computer and Consulting GmbH
München
+491725328003