Using Exploratory Data Analysis to Discover Patterns
summary
Develops an EDA pipeline that extracts features from images and documents using pre‑trained deep nets (ImageNet/Caffe model zoo) and text features (tf‑idf), turns image collections into numpy arrays via a scikit‑learn style wrapper, and visualizes them with PCA/t‑SNE and montage-style plots. Builds PyImagePlot and a ski‑caffe wrapper, processes 10,000 images interactively, spots a July 3 fire event in the camera stream, and finds a persistent ~2× he/she imbalance in campus newspapers.
AI-written summary · about
chapters (9)
- 0:00Exploratory data analysis intro
- 3:57Classic EDA pipeline limits
- 8:37Visualizing high-dimensional data
- 12:31Twitter image time and hue
- 17:47Image feature extraction deep learning
- 23:05Wildfire detection case study
- 26:37Document analysis text features
- 30:06Gender pronoun analysis
- 35:05Summary
related talks
37mData Science for Digital Humanities: Extracting meaning from Images and Text
24mExploratory Data Analysis (EDA) and Visualization Techniques..
40mData-centric exploration using intake, dask, hvplot, and more
01h 00mData, visualization, and designing AI
01h 48mApplied Data Science
24mThe TL;DR of EDA
01h 29mComputer Vision: An (Un?)Expected Journey, with Keras and Tensorflow
38mTransforming Data to Unlock Its Latent Value