I've captured Twitter data of Palm Springs area since the beginning of the Esri Dev Summit
on Sunday 8th of March until Wednesday 11th of March for our presentation. I developed an application with the results and is available here:
The main features of the application are:
- a heat map of the tweets: here you can appreciate most of the activity in the Palm Springs Convention Center where the Esri Dev Summit is taking place and also in the Tennis Garden of Indian Wells where the BNP Paribas Open 2015 of tennis is taking place.
- Classification of tweets: Blue tweets are tweets talking about Esri Dev Summit, Green about the Tennis tournament and red about any other topic.
- When you click on the tweet you get the tweets text, the timestamp and the user.
It is a very simple use case a much more can be done.
In order to create this application I used IBM SPSS Modeler 17
, Esri ArcGIS
and Esri Web Application Builder that I am going to explain now with more detail.
To get the Twitter data I used IBM SPSS Modeler 17 I used a new Extension developed with R to connect to the Twitter Streaming API
. This extension is not yet available. Using this API I can choose a bounding box of the area I am interested, and I selected to get tweets from Palm Springs area using the Bounding Box website, very useful for this:
Then I am doing very basic text extraction to classify the tweets:
Esri_tag--> issubstring("Esri", text) or issubstring("devsummit", text) or issubstring("DevSummit", text) or issubstring("GIS", text) or issubstring("esri", text) or issubstring("Geo", text)
Tennis_tag--> issubstring("Indian Wells", text) or issubstring("BNPPARIBASOPEN", text) or issubstring("#BNPPO15", text) or issubstring("#IndianWells", text) or issubstring("Bnp", text) or issubstring("bnp", text) or issubstring("tennis", text) or issubstring("bnpparibasopen", text)
Obviously this is a joke compared to the things you can do using the IBM SPSS Text Analytics features within IBM SPSS Modeler. But for this simple demo it was enough.
Then using another new SPSS Extension that I developed with R I am pushing the data back to Esri ArcGIS online.
All the results of my analysis are now in my ArcGIS online. In this data I have the location of each of the tweets and the category. Once I have the data in ArcGIS online I can start using all the powerful features of Esri to plot GeoSpatial data and create Map Based applications.
In this case to share the application I used the Esri Web Application builder
. This is an amazing online tool to create beautiful map applications within minutes and share them with everybody. Learn more about it here: