Hi Sarthak,
I played a bit with your data, and this is some findings.
- I seems you are trying to draw your data on the map twice. In Cognos as regions and as points, and in Tableau in blue and green. I am not sure what you are trying to achieve by this. It helps if you could explain what your map should show.
- If you want 2 point layers on the map, the only way in Cognos to do that is to georeference your spreadsheet, by adding a lat+lon column to it. There are online tools to help you with that. (I found, but did not test
https://www.geocod.io/upload/ )
About the data (which should have effect in Tableau as well):
- There are 2 postal codes in your data that Cognos/Mapbox can not find. They seem to be valid postcodes though. I am not sure what process IBM has for making these postcodes mapable and I can not find where Mapbox is pulling their postcode information from. The postcodes are E0L and E0J
- There are several Ambiguous Regions. I checked one, Moncton. What is strange about Moncton is that there really is only 1 Moncton (in OpenStreetMap), so I do not understand how it could be ambiguous. As this is a region, I presume it is coming from the Mapbox Boundaries layer. You can find more about the
source of that layer here Again, you have to work with IBM and Mapbox to figure out why Moncton (and the other regions) are ambiguous.
- There are also a few locations Unrecognised. I checked Deersdale. Now, I presume this layer is based on the Mapbox Street v8 layer and the way that layer calculates the attribute label_name is based on items in OpenStreetMap from
this list. For Deersdale the type was 'locality', as Deersdale seems to be not much more than a mine and the buildings around it. 'locality' is not included as a valid placetype. So, as an experiment, I changed the type of place in the OpenStreetMap database to 'hamlet' which might not be true (i do no know if there are people living in Deersdale), but should have the effect that Deersdale will appear on the Cognos *and* Tableau map in the near future. If so, you could do the same for the other localities that you are trying to map. Please keep in mind that the OSM database is a public database, and your mapping should reflect the real world and
hamlets have inhabitants, and
localities have not.
For those of you who are wondering about Google. Please compare
OpenStreetMap's view of Deersdale, with
Google's, and you know why so many companies prefer the former.
Maybe someone from IBM can enlighten us on how they expect their clients to enhance the data behind the maps.
Hope this helps a bit,
------------------------------
Reinier Battenberg
https://github.com/CognosExt------------------------------
Original Message:
Sent: Tue January 21, 2020 02:09 PM
From: Sarthak Dalmia
Subject: Unable to create Geo spatial dashboard in Cognos
Hello Kasun,
Thank you for your reply. I am attaching the data source excel and the maps that I created using Cognos. I also used Tableau. I wanted to know if the screenshots in Cognos are the best that we can do in Cognos. In the word document, I have also pasted what I did in Tableau. Could you please check and let me know?
Thanks,
Sarthak
------------------------------
Sarthak Dalmia
Original Message:
Sent: Tue January 14, 2020 01:46 PM
From: KASUN ATTANAPOLA
Subject: Unable to create Geo spatial dashboard in Cognos
Hi Sarthak,
Looking at the attached ticket, I see a couple of ways we can improve mapping accuracy here.
1. For the points layer, we have just the the city data added and Fairfield is marked as an ambiguous region. This just means that the mapping database has identified multiple "Fairfields" and has taken a best guess and finding the right location. You can disambiguate instantly by qualifying the City with the next level parent region (in this case State). Now the system will try and resolve cities with with the given "state" context. As shown in the image below, the ambiguity is now resolved since the system understands in this case "Fairfield" is located in the State of "Maine"
------------------------------
KASUN ATTANAPOLA
Original Message:
Sent: Mon January 13, 2020 04:13 PM
From: Sarthak Dalmia
Subject: Unable to create Geo spatial dashboard in Cognos
I am trying to create a geospatial dashboard in Cognos. I did check the sample data on the Cognos website and used the same format for creating a dashboard. Below is the data source that I used. I have the data (city/province/ country) for both the USA and Canada. I am attaching the data source and the error I am getting. Can anyone help me identify the data format and the column fields that should be used to map the data? I am also attaching the final file that I want to upload in Cognos. Could anyone help me in identifying the data format that Cognos will be able to read?
| Country | State | City | Postal Code |
| United States | Maine | Bangor | 70001 |
| United States | Maine | Fairfield | 70005 |
| Canada | Saint John | Saint John | E2A |
------------------------------
Sarthak Dalmia
------------------------------
#CognosAnalyticswithWatson