dmr / ovc-2016-videos

Code for visualizing videos from OpenVis Conf 2016

ovc-2016-videos

Code for visualizing videos from OpenVis Conf 2016.

Analysis

The text analysis was done in Python (tested with Python 3.5) and is stored in the analysis directory. To run the scripts, simply run python <script.py> (e.g., python analyze_top_terms.py)

Script	Description
analyze_bigrams.py	Finds the top 50 bigrams in the transcripts for each talk
analyze_tifdf.py	Finds the top 50 terms in the transcripts for each talk by TF-IDF score
analyze_top_terms.py	Finds the top 50 terms and top 50 bigrams from each talk, maps them to a similar score and matches them with timestamps. This is the main script used by the visualization.
analyze_trigrams.py	Finds the top 50 trigrams in the transcripts for each talk
match_timestamps.py	Matches terms and bigrams to timestamps in the transcripts

Visualization

The visualization code is in the app directory. The vis was built using React and D3.

Set up

1. In the app directory, install dependencies with npm install
2. To get the development server with hot-reloading started, run grunt
3. Open a browser to http://localhost:8080

Building

1. To build files for production, run grunt deploy

Data

The transcripts are available both with timestamps and without in the data directory.

Contact

For any questions, please email [e-mail].