Improving Building Accessibility using Machine learning

Check out our website here!

This repository contains the notebook and data to accompany the implementation. This implementation was also a whack at utilizing different technology, Wolfram Mathematica.

Downloading Data

Before running the notebook, you'll want to download all the data you'll be using located under yhack_data.zip.

• the folder contains multiple .tsv files for both walking on flat ground and walking up the stairs.
• stairsdown denotes the raw dataset for walking down the stairs
• stairsup denotes the raw dataset for walking up the stairs
• data.tsv denotes the raw dataset for walking on flat ground

To clone the repository:

git clone https://github.com/kmualim/yhack

Unzip the data file provided and run the preproceessing script on the corresponding raw datasets

bash process.bash

• process.bash proceeds to isolate the x,y,z coordinates of the given activity for input into our algorithm.
• the output consists of (stair/walk)(x/y/z).csv files and these x,y,z coordinates were then subsequently concatenated to produce a spectogram
  

Figure 1: Spectrogram of "Walk"

Requirements and Installation

In order to run the notebook, you'll need

• Wolfram Mathematica