Data Visualization with PyScript, Pandas, Colab and more AI
Data Visualization on the Web with Plotly and PyScript
(Medium article - no paywall with this link)
Put your data visualisations on the web with HTML templates — no server-side code required. A new article that shows you how to (as far as possible) avoid the dreaded HTML when creating PyScript data visualizations.
Related articles: Hello PyScript, What’s new in PyScript in 2023.
10 Minutes to Pandas in Colab
I’m sure we are all familiar with, at least, the existence of Pandas - it’s a popular open source Python package for data science and you can see it being used everywhere.
10 Minutes to Pandas is an introductory tutorial provided by the Pandas project and now it has been ported to Google’s Colab (the online notebook app) so you learn the basics of Pandas with an interactive version the tutorial. You can find it here.
A Close Look at Colab’s new updates and enhancements
(Medium article - paywall)
Staying on the theme of Colab, Parul Pandey has written this update to using Colab and describes the features Smart Data Pasting from Google Sheets, Automated Plot Generation from Pandas DataFrames, and Interactive DataFrames among others. If you have been keeping up with Colab developments this is definitely worth a read.
Low Code AI
I got a notification of this new book from O’Reilly in my email the other day and although I haven’t quite got around to looking at it yet, I’ve put it on my reading list and thought I’d share it. It looks like it could be interesting.
To quote O’Reilly’s press release, Low Code AI gives you “three problem-focused ways to learn no-code ML using AutoML, low-code using BigQuery ML, and custom code using scikit-learn and Keras. In each case, you'll learn key ML concepts by using real-world datasets with realistic problems”.
You can get a preview of the first couple of chapters from the links above.
Computational Agents Exhibit Believable Humanlike Behavior
This is a fascinating read from Stanford University’s Human-Centred Artificial Intelligence (HAI). It’s a report on research conducted by Stanford PhD student, Joon Sung Park and collaborators, who have created a simulation populated by agents that exhibit human-like behaviour. But the behaviour is not a result of programming by coders, rather they are given a short biography “consisting of a name, age, job, family, interests, and a few habits”, and their behaviour is driven by an LLM similar to ChatGPT.
HAI say, “Generative agents rely on a large language model to remember their interactions, build relationships, and plan coordinated events, with implications for both gaming and social science”.
In an experiment where these agents were ‘interviewed’ by researchers and their responses compared with those from real people, human evaluators rated the response from the AI agents as exhibiting more believably human-like behaviour than the real human responders.
This article contains affiliate links. If you make a purchase I may receive a small commission but it will not affect the price that you pay.