Intro to Machine Learning
From definitions to examples
John Alexis Guerra Gómez
🇨🇴@guerravis
🇺🇸@duto_guerra
https://johnguerra.co/viz/introToML
Use spacebar and the arrows to advance slides
Data Science
What is it?
Data science is an inter-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from structured and unstructured data
Wikipedia
... to work effectively with heterogeneous, real-world data and to extract insights from the data using the latest tools and analytical methods.
UC Berkeley MIDS program brochure
However
When you search online this is what you see
Google Trends
Data Science is way more than Machine Learning!
The purpose of visualization is insight, not pictures
The purpose of data analytics is insight, not (just) models
What part of ML do you want to use?
Prediction
How to use it (prediction)?
- Op1. Use models (apis) trained by Google, Amazon, IBM or Microsoft
- Op2. Download models pretrained by others and run them yourself
What can you use it for?
- Photos 🖼
- Videos 📹
- Document/Text Processing 📃
- Speech 👄👂🏼
- Structured data 💾?
Photos 🖼 and Videos 📹
What can I detect on photos 🖼?
- Objects 🐈 🐕 🍎
- Faces 👱🏽♂️👱♀️
- Celebrities 🍾
- Landmarks 🗼
- Text in images 🗼
Video 📹 is about the same but on streaming
How can I use it?
Develop locally
Pose Detection
https://johnguerra.co/viz/mlPose/Object Detection
https://johnguerra.co/viz/mlObject/How can I use it?
- Big companies 👉 ML as a service
- You pay per request 💵 ~US$1.0 per every 1000 requests
Demos
Document/Text processing 📃
What can I do with documents 📃?
- OCR 🖼 → 🔤
- Sentiment analysis 😆😡
- Topic extraction 🟡🟠🟣
- Entities detection
- Political Affiliation? 👔🎉
- Psychological Profile?
Demos
Speech 👄👂🏼
What can I do with Speech 👄👂🏼?
- Speech recognition 👂🏼
- Speech generation 👄
Data Analysis
Let's compare them with a real world example
How is Rappi doing on Twitter?
- 30k tweets in a week of 2019
Approach 1
😡😠😒😐😏😃🥰?
- Machine learning 🎩! ???
- Detects sentiment ! ???
I hired a data 🐒 (might be me)
Analyzed 180 tweets
- 😡😠😒😐😏😃🥰
Here are some of them
😐 -10%
😡 -80%
🥰 80%
😐 -10%
😐 -20%
🥰 90%
😒 -40%
😒 -30%
Would you hire this data 🐒?
Well, actually
- It wasn't a data 🐒
- It was a 💻
- Would you use it?
Well, actually, actually
- It wasn't just any computer
- It was ML as a service by Google
- It costs $$$
Will you trust it?
I don't
- Spanish is just too complicated
- These models are trained with 🇪🇸 examples
Approach 2
Approach 3
It's up to you!
- Interactivity 👉 Ask questions
- Slice and dice
- Overview first, Zoom/Filter, then details on demand
Rappi Dashboard Link 😉
¡No coma Machine Learning, coma 🍖!
My insights toolset?
Training
Tensorflow Models
https://www.tensorflow.org/js/modelsTensorflow.js
Multivariate Data?
-> Dimensionality Reduction + Clustering
MLExplore.js
Visual Analytics Applications
- Present information
- Discover insights
- Preprocess + Explore
- Open the black box
Present information
- Prediscovered insight
- General public
Discover insights
- For experts
- Navigate the data
Preprocess
- Understand new data
- Cleanup
Cleanup
Explore
Open the black box
MLExplore.js
- Interpret and interact with TSNE+Kmeans