Mini-Project: A Flask App That Filters

In this mini-project assignment, you will build a web-browser interface for a dataset. That is, your web application will:

• Take in user input about how to filter/search the dataset
• Pass in those user inputs (which are serialized as a dictionary) to Python functions that do the work of reading data files, sorting, filtering, and returning the results.
• Pass the results into an HTML template page.

Lessons can be found here: https://github.com/datademofun/flask-data-querystrings

Example dataset and app

Check out the College Scoreboard app, as presented by the U.S. government:

https://collegescorecard.ed.gov/

Here's the source data: https://collegescorecard.ed.gov/data/

Instead of downloading the entire ZIP file of data, pull in these two CSVs:

• Scorecard data
• Post-school earnings

Your own school app should allow the user to search, filter, and rank in a way different than what the College Scorecard provides. You may even want to simplify the approach.

The two data tables allow you to find post-school earnings by given school (if you join the appropriate fields). Don't think of showing all the information, just the most interesting information.

If you do a data join while creating the listing view, you could give the user the option to rank schools by post-earnings salary, among other things.

You can pick your own dataset, but it has to be something in which the number of data fields is numerous and interesting enough that there would be a reason to have a separate listing and detail view.

For example, New York restaurant health inspections:

The raw dataset is on Socrata as a flat table (instead of separate tables for inspections and restaurants, it's all one table with redundant data).

• Check out NYC's official website for the data.
• And then check out the New York Times's riff on that data
• You can google around for other uses of the data by citizen data journalists.

Functional requirements

Build a data web application that takes in user input for:

• an attribute to filter by, e.g. (row['gender'] == 'F' or 'Dan' in row['name'])
• an attribute to sort by, e.g., sort by the 'birthdate' attribute of each row, in either ascending or descending order.

The main view of the app should include a web form that has these filters.

The results view should show a web-friendly listing of the source dataset, and each list item should include an internal link to a details viewpoint (a la the detail.html template in the First News App tutorial).

The detail view should include more detailed data about the selected data row, including data that you did not deem fit to include in the listing view.

Deliverables

For Monday, May 16

• A Github repo named flask-filter-app that contains (or will eventually contain) the data and code to run and deploy your app. It should be separate from cj-2016, i.e. at github.com/yourname/flask-filter-app, not github.com/yourname/flask-filter-app
• A README.md file at the root of your flask-filter-app folder. Using Markdown syntax
  • Describe the two (or more, if you choose) views of your app:
    • What is shown on the main listing page
    • What is shown on the details page
  • Describe the filter options – e.g. full text search of name, sort by a date field, a checkbox for 'M' or 'F', you will make available to the user.
  • List three actual live data applications (such as ones produced by the government, or by news sites) that inspired you.
• Have a more-or-less complete foo.py file
  • You aren't required to have an actual complete web application, or even one that works at all. However, you do need to have the functions that work with the data in a separate foo.py file.
  • At the very least, I should be able to download your repo, run ipython and call foo.print_record_count() and get an answer. In this particular app example, look at how app.py calls in foo.py

For Wednesday, May 18

• a URL on which your Heroku-deployed app lives
• An app in which all of the endpoints actually work