Geospatial Datasets

[Atabarezz]

After the intense interest shown to the community topic Available "Big data sets" over the internet...  (since ‎22-11-2015)

57REPLIES

20829VIEWS

I'm now trying to split that into more meaningful, more focued sections. So that people can get an understanding of

1. industry specific data and
2. business problems that can be solved easily with Alteryx...

This specific section is on #Geospatial datasources, maps, road network data, traffic data, footfall data, points of Interest, DEM (digital elevation maps)

[r4upadhye]

Hey, Im interested in geospatial data analysis, I had a mini module for analyzing predefined boundaries and polyshapes for one of our client. It was for a Polish Railway Company though.

Would be interested in contributing to your version too.

[Atabarezz · Accepted solution]

POI datasets are extremely useful to enrich your customer databases with the help of Find Nearest Neighbors tool.

But these datasets are not easy to find and they may be expensive if you cover a vast geography.

The openstreetmap community is going great;

http://poi-osm.ekibox.net/

Here is an address for you to download POI in KML format...

[marco_zara · Accepted solution]

That info is pretty old (2014!), we can do better with Alteryx!

I successfully managed to extract all OpenStreetMap features from their XML files on a centroid basis (but we can easily rely on geofrabrik datasets for polygons), if you're interested I can dig up the workflows...

[Atabarezz]

Exactly! you may be aware there are some spatial data set for alteryx around 15K -20K USD/year, pretty expensive for an SME(small medium enterprise)...

With this data plus some contribution from Alteryx people to the openstreetmaps

we can not only cover North America, Canada and Europe but the World...

Great opportunity...

[marco_zara]

I can whip up a tutorial in the weekend, just be prepared for extremely long waiting times (I mean days or more) on large datasets, IIRC Italy took me around 3 days to convert.

The good side is that OSM potentially has thousands of features for a single point, the bad side is that any point will have usually up to a couple dozen actually populated.

[estherb47 · Accepted solution]

Here is some interesting information on US geospatial: https://catalog.data.gov/dataset?tags=cities

And, while I haven't used it yet, this looks promising! https://github.com/kelvins/US-Cities-Database

Cheers!

Esther

[marco_zara · Accepted solution]

I managed to recreate my OSM feature extraction datasets, however on larger areas I think the sheer number of columns breaks the crosstab tool...

Base file is here: https://download.geofabrik.de/europe/italy/nord-ovest-latest.osm.bz2

Unzip it and run the two workflows, any feedback is extremely welcome!

Barring a few small issues this can be used along with OSM shapefiles to obtain some extremely powerful geographical profiling.

With a few community refinements we might turn this into something useful!

[marco_zara]

Interesting, crosstab seems to get issues at around 600-650 fields.

It looses track of already created ones and detects non-existing duplicates, which is an issue.

[Atabarezz]

I work with over 1000 columns as well and that happens... Great point to raise to the support teams...