Geographic Hierarchies - How to get list of valid Region and Sub-Region values for the UK?

[kevin_small]

I would like to get a list of UK regions and sub-regions that Lumira will accept as valid values for a geographic hierarchy.

The reason is that I have some source data that has geographical information held as UK local authority names (these are the local councils).  Lumira does not recognise these, which is understandable, and does not propose anything sensible for me to replace them with.  There are not so many UK local authorities, so I can convert these by hand, but I need to know what the valid target values are for Region and Sub-region.

The documentation says "Geography hierarchies can only be created on columns containing values compatible with geography data values in the NAVTEQ database used by SAP Lumira".  I've looked online for NAVTEQ master data lists, but I cannot find anything usable.

Does anyone know how I can see lists of valid UK regions and sub-regions that Lumira will recognise?

[Henry_Banks · Accepted solution]

Hi Kevin,

i think your best options here is to use a geographic dataset which has all the latitude & longitude coordinates for the locations you want to map, down to a level of detail 'postcode' for example.

the limitation with the inbuilt libraries is that the population has to be greater than 100K for it to be recognised.  For example, our own dear Walldorf isn't detected when using 'names' either!

Regards,

H

[kevin_small · Accepted solution]

Hi Henry,

Thanks, and can you clarify something for me?  When you say "the population has to be 100k" to be recognised, do you mean 100k unique geographic values are required, or could it be 100k transaction records sharing only a few unique geographic values?

The reason is, I have a few million transaction records.  On each transaction record I have a latitude, longitude pair as well as a field holding the local authority.  I have already tried using the latitude, longitude pair as the basis of a geo-hierarchy in Lumira but performance was terrible.

I would like to get the backend (HANA in this case) to do the heavy lifting of aggregating the data up to a manageable number of records at region/sub-region level before handing the resultset to Lumira.

Thanks,

Kevin.

[Former Member · Accepted solution]

Kevin,

Henry is referring to the actual population of the city/town/village, and how it must be at least 100k to be in the database, and thus recognized.

Cheers,

Bijan

[Henry_Banks · Accepted solution]

Hi,

What Bijan confirms is correct for 'names' detection. There are some tutorials here

With regards to your performance concern: you've got 2 HANA connection types, online, and download.

the 'download' connection does actually import the data locally into the sybase instance on your machine. You typically get a warning after approx 30 million cells (rows x columns).

For example, I just refreshed 2M rows x 10 columns (20 million cells), over a very poor VPN connection,  in just over 3 minutes.

The 'online' connection is quicker. It's not a 'limitless' refresh size, but i know it has been pushed to over 125 Million rows of data and over 40 columns.

Regards,

H

[kevin_small · Accepted solution]

That's great thanks Henry and Bijan, I will take a look at those links.

I did play with the download option to keep the data in sybase locally, and you're right once you've got the data downloaded performance is fine.  I was wondering, though, about how to cope with scenarios where there are a variety of consumers of the data with different front-end tools.  In these cases I might want to correctly enrich the geo data on the backend first, and send a single, consistent set of data to all clients.  I suppose in that case I'm not using Lumira for its geo-enriching capabilities, but just for it's geo-visualisation capabilities.

[Henry_Banks · Accepted solution]

Yes that sounds right .

in terms of personas - You could take the 'desktop clien't as being the power-user/designer's tool: to cleanse, enrich, merge and manipulate your data.

then, the consolidated dataset is then shared to Cloud https://cloud.saplumira.com/  where your end-consumers can discover said dataset, and explore any visualisations/stories you may have composed & exported.

regards,

H

[kevin_small · Accepted solution]

Thanks Henry, that makes sense.

[kevin_small · Accepted solution]

Hi again Henry.  I've gone through those links thanks and had some success, thanks for that.

I'd like to make a closing point:

I agree the scenario you suggested of self-service enrichment and user-publishing to Lumira Cloud is one valid use-case.

However, there is another use-case that would benefit from valid lists of regions/sub-regions/cities.  In an enterprise deployment, it is certain that some customers will want the cleansing and enrichment of very granular data to be done centrally by IT, and then the common aggregated resultset handed out to many different systems and clients - just one of which might be Lumira.

Without knowing the valid lists of geo master data that Lumira will recognise it would be a difficult task to standarise the results such that Lumira can report on the same data as the other systems and clients.