seaborn VS Altair

Below are the key insights. If you want to see the Python code I used to do this analysis and generate the charts using Seaborn, you can find my full analysis Jupyter notebook on my Github repo here: Tip Analysis.ipynb

seaborn

Additionally, Seaborn (https://seaborn.pydata.org/) is a great mention for people that want to use Matplotlib with better default aesthetics, amongst other conveniences:

"Seaborn is a Python data visualization library based on matplotlib. It provides a high-level interface for drawing attractive and informative statistical graphics."

Seaborn: built on top of matplotlib, adds a number of functions to make common statistical visualizations easier to generate.

pandas - The standard data analysis and manipulation tool numpy - scientific computing library seaborn - statistical data visualization sklearn - basic machine learning and predictive analysis CausalML - a suite of uplift modeling and causal inference methods PyTorch - professional deep learning framework PivotTablejs - Drag’n’drop Pivot Tables and Charts for Jupyter/IPython Notebook LazyPredict - build and work with and compare multiple models phidata - Build AI Assistants with memory, knowledge and tools. Lux - automates visualization and data analysis pycaret - low-code machine learning library. really nice Cleanlab - for when you are working with messy data drawdata - draw a dataset from inside Jupyter pyforest - lazy import popular data science libs streamlit - simple ui builder, useful for demonstrating ML results

How to Accomplish: Utilize visualization libraries like Matplotlib, Seaborn, or Plotly in Python to create histograms, scatter plots, and bar charts. For image data, use tools that visualize images alongside their labels to check for labeling accuracy. For structured data, correlation matrices and pair plots can be highly informative.

If you are doing data analysis I don't think any of the 3 pieces of software you mentioned are going to be that helpful.

I see these products as tools for data visualization and reporting i.e. presenting prepared datasets to users in a visually appealing way. They aren't as well suited for serious analytics.

I can't comment on Superset or Tableau but I am familiar with Power BI (it has been rolled out across my org), the type of statistics you can do with it are fairly rudimentary. If you need to do any thing beyond summarizing (counts, averages, min, max etc). It is not particularly easy.

For data analysis I use SAS or R. This software allows you do things like multivariate regression, timeseries forecasting, PCA, Cluster analysis etc. There is also plotting capability.

Both these products are kind of old school, I've been using them since early 2000's, the "new school" seems to be Python. Pretty much all the recent data science people in my organization use Python. Particularly Pandas and libraries like Seaborn (https://seaborn.pydata.org/).

The "power" users of Power BI in my organization tend to be finance/HR people for use cases like drill down into cost figures or Interactively presenting KPI's and other headline figures to management things like that.

It's referring to the seaborn library (https://seaborn.pydata.org/), a Python library for data visualization (built on top of matplotlib).

While it’s not perfect and it’s not ggplot2, Seaborn is definitely a big improvement over bare matplotlib. You can still use matplotlib to modify the plots it spits out if you want to but the defaults are pretty good most of the time.

https://seaborn.pydata.org/

Altair uses Grammar of Graphics principles:

It's compatible with a wide range of data libraries, including Pandas, NumPy, and Altair. Streamlit integrates with all the latest tools in generative AI, such as any LLM, vector database, or various AI frameworks like LangChain, LlamaIndex, or Weights & Biases. Streamlit’s chat elements make it especially easy to interact with AI so you can build chatbots that “talk to your data.”

I like Vega-Lite: https://vega.github.io/vega-lite/

It’s built by folks from the same lab as D3, but designed as “a higher-level visual specification language on top of D3” [https://vega.github.io/vega/about/vega-and-d3/]

My favorite way to prototype a dashboard is to use Streamlit to lay things out and serve it and then use Altair [https://altair-viz.github.io/] to generate the Vega-Lite plots in Python. Then if you need to move to something besides Python to productionize, you can produce the same Vega-Lite definitions using the framework of your choice.

Feel free to open an issue to let us know which parts of the documentation you find obscure and if you have suggestions for how to improve them. We did a larger overhaul a few months back and are always open to feedback on how to improve it further! https://altair-viz.github.io/

(disclaimer: I'm a co-maintainer of Altair)

Vega-Altair is pretty great as well. It uses a grammar of graphics that’s slightly different from ggplot, but has most of the same advantages.

https://altair-viz.github.io/

Altair - Declarative statistical visualization library for Python.

Github: Altair