台北市都市更新與土地利用

Renewing Taipei

This project was done in the context of EPFL | Media & Design Lab | Personal Interaction Studio 2018

Course Tutors:

• Immanuel Koh
• Jeffrey Huang

In this repository, we collect data from multiple sources regarding renewal areas in Taipei City starting year 2012.

Our data is obtained thanks to

• Taipei Historical Maps
• 地號 GeoJSON API for GeoJson Files
• Taipei Urabn Development Department for Illegal Rooftops
• Taipei Urban Regeneration Office for Renewal Areas
• Taiwan National Development Council for Land Prices

Screenshot & Demo

How to run it

1. In the cloned directory, use python -m http.server to run the server
2. http://localhost:PORT_NUM/Website/index.html, and here it goes!

Folders

Data

In this folder, we organize our obtained and cleaned data. The final result concatenating everything is Final.csv

Particular files of interest:

• Land Prices 201X.csv: Land price data for year 201X
• Output.csv: List of Illegal Rooftops and their coordinates

IPython

This folder contains jupyter notebook files where we experimented with different methods.

• GreenFinder.ipynb: This notebook shows how we calculate the vegetation indices
• TSNE.ipynb: This notebook shows how we built our TSNE posters

N.B.: The images are not uploaded to github due to their number and size. They can be recovered by using Scripts\SnapRenewalAreas in QGIS

Latex Stuff

This folder contains .tex files we used to generate our report for the course

OCR

This folder contains the scripts used to download and parse renewal areas data. It also contains the data in different subfolders:

• GeoJson: GeoJson data files for each renewal area
• LandNumber: Subdivisions of each renewal area
• Text: Raw uncleaned text obtained after OCR on the scans

Data was collected with the following pipeline:

• Crawler/Renewal Pdf Crawler.ipynb: Retrieves renewal pdfs from Taipei Urban Regeneration Office
• extract.py: Takes only the first few pages from the raw pdf to speed up the OCR process
• OCR.py: Gets text from the extracted pdf using Google Drive API
• parser.py: Basic cleans-up and parses land numbers from the text we obtained
• getJson.py: Gets GeoJson files using 地號 GeoJSON API

Rooftop Address

Raw data for the illegal rooftop locations (District/Street Name)

Scripts

Python scripts used in QGIS in order to analyse and generate our final.csv file as well as snapshots at different zoom levels for the satellite and infrared imagery:

• AddressToLatLon.py: Transforms the raw illegal rooftop address into usable (lat,lon) coordinates
• DrawOnMap.py: Reads the illegal rooftops and renewal area data and displays them on the map
• GenerateRenewalsQGIS.py: Generates the first part of the information about renewal areas relating to illegal rooftops
• GenerateRenewalPython.py: Generates the second part of the information abotu renewal areas relating to land prices, vegetation and dates
• SnapIllegalRooftops.py: Takes snapshots of all the illegal rooftops and stores them in a specified folder
• SnapRenewalAreas.py: Takes snapshots of all the renewal areas and stores them in a specified folder
• Utils.py: Commonly used functions and variables

Website

This folder contains the website that demonstrates our result.

• index.html
• js/ui.js
• css/columns.css

Icons thanks to

• Gregor Cresnar
• Freepik
• Smartline
• Smashicons

from Flaticon