MapMaker

(This is the publicly available documentation for MapMaker, which is a closed source application under copyright by Pedal Technologies, Inc. The Eurasianet logo and branding are property of Eurasianet)

MapMaker is a crossplatform application for producing publication-quality maps of countries in Central Asia and the South Caucasus. Pedal Technologies was commissioned by Eurasianet to design and build MapMaker, and it has been used to produce maps for several of their articles (such as here and here). Eurasianet is a media organization based at Columbia University’s Harriman Institute, one of the leading centers in North America of scholarship on Eurasia.

I led development on this project, which included the following tasks:

• Designing UI/UX
• Building the infrastructure to generate base maps for each country (using matplotlib and data from Natural Earth)
• Designing and implementing an efficient algorithm to place markers, labels, and connectors on the map automatically, without overlapping with country borders on the base map
• Creating shapefiles for disputed regions that do not have publicly available geospatial data (using QGIS)
• Packaging the Flask application using Pyinstaller to run seamlessly on Windows, Mac, and Linux

Features

Provides support for eight countries in Central Asia and the South Caucasus

• The user can input the name of a village/town/city target in one of the following countries:
  • Uzbekistan
  • Turkmenistan
  • Tajikistan
  • Kyrgyzstan
  • Kazakhstan
  • Georgia
  • Azerbaijan
  • Armenia

Produce maps in English or Russian

Eurasianet reports on news to a diverse audience of readers, so the user can produce maps in either of these two languages with ease

Autocomplete for target names

As the user inputs the name of a location, some autocomplete options are displayed so that the user can quickly find the name of the target

Accurate geocoding for quick use

MapMaker harnesses the power of OpenCage Geocoder to retrieve accurate coordinates for targets

User-friendly error handling

If the results from geocoding are unsatisfactory (generally when the target is not well known), the user is prompted to manually input coordinates for the target

Efficient algorithm for feature placement

• The most challenging technical problem in this project was developing an efficient algorithm to place multiple labels, markers, and connectors on the base maps without creating overlap between any map features (See Automatic label placement)

• The greedy algorithm iterates outward from the geographic coordinates of each feature and checks if the bounding boxes of the label and its connectors overlap with those of any other features.

• In the demonstration below, I have made the bounding boxes visible to illustrate how the algorithm finds the locally optimal location for each label:

• This is what the map looks like when completed:

How to use

1. Select language
2. Input target name or coordinates (If language is Russian, input target name in English as well as Russian. The former is required for geocoding, and the latter for the map label.)
3. Click 'Create Map'
4. If the map preview is satisfactory, click 'Save' to store the map as a .png file
5. Otherwise, repeat steps 2 and 3 until satisfied