Skip to content

Latest commit

 

History

History
29 lines (22 loc) · 2.31 KB

about.md

File metadata and controls

29 lines (22 loc) · 2.31 KB
Raw

About this data portal

This project was built and is maintained by Anthony Moser to make it easier to search Chicago's data portal (and others like it).

It's a fairly simple Python app that uses sodapy to access the data portal and Streamlit to create this web interface. You can check out the repo at http://www.github.com/anthonymoser/better_data_portal

Notes on usage

  1. Maps and "derived views" are excluded by default. The app automatically excludes data sets that are categorized as maps, because keyword search doesn't work on maps. It also excludes "derived views" because they're just filtered copies of a full data set that will still be searched. For example, "Current Employee Names, Salaries And Positions - Full Time" is just a subset of "Current Employee Names, Salaries And Positions." To include these (or see them in the list of available data sets), check "Include subsets"
  2. You may get some partial matches. Keywords are automatically treated as having quotes around them (e.g. My Search will be treated as "My Search"). Despite this, some data sets will include partial matches, which can be false positives. For example, a search for "4010 S Damen" returns Divvy bike rides with the user_id "4010".
  3. I don't store anything. None of your searches are logged or results are stored by this application.
  4. Datasets are grouped automatically. To make the list of selectable data sets more managable, they are grouped by whatever comes before the first dash. For example, "311 Service Requests - Dashboard" and "311 Service Requests - Abandoned Vehicles - No Duplicates" will be combined into a single checkbox for "311 Service Requests." When you choose a group, you will search all sets in the group, which will be individually listed in your "Selected Data Sets."
  5. It's okay to just search everything. Part of the power of this approach is that you don't know where something will turn up.
  6. Leave off street designation (RD, AVE, etc). Some datasets store the address separately from the street type. For To avoid accidentally excluding results, "2055 W Pershing" is better than "2055 W Pershing RD"

Where can I find Socrata data portals?

Many Socrata powered datasets can be found at http://www.opendatanetwork.com Just search for the place or topic and look for the "datasets" results.