Practical Business Python

Taking care of business, one python script at a time

Newsletter Number 7 - It’s a bit meta

Sent on Sat 02 March 2019

In this newsletter, we will take a quick tour through some of the latest posts on PB Python and some other interesting updates in the Python community.

Around the site

  • Finding Natural Breaks was an interesting article to write. Prior to researching the article, I had never heard of the term “Natural Breaks” or the Fisher-Jenks algorithm. This is an useful algorithm to keep in mind when binning data sets.
  • Interactive Dashboards is another guest post from Duarte O.Carmo. I appreciated this article because it brought together many concepts including using a basic reddit API, NLP and visualizing data with Plotly Express. Finally, he covered how to deploy this solution. Check it out if you have not seen it.
  • Using Markdown to Create Responsive HTML Emails is a bit meta. I’m writing this newsletter using the work flow described in the article. This may be a bit specialized but was fun to write and hopefully is useful to others. This step also completes my migration away from Mailchimp.
  • Record Linking descibes some of the big challenges I have personally experienced trying to link records together without a common unique record identifier. In addition to the tools I discussed, the comments include some other additional ideas that I may investigate in future articles.

Other news

  • Many of you have likely seen news about the Pandas 1.0 release. I have not tried out many of the new features related to the new data types- I want to let it settle out before I use it extensively. However, I am definitely excited about the new string type. I have been posting a few pandas tips on twitter and will likely continue to do so as I explore more of the 1.0 changes.
  • A while back, I saw this new Python training series from Microsoft. It has very high production values and could be useful to those looking for a video intro. The code is also available.
  • Speaking of training, the NSA has released its curriculum for python training. Unfortunately it is only in PDF but it is complete and very well done. Not surprisingly, it includes some in-depth content on HTTPS and PKI. You can definitely read between the lines and understand more about how intelligence organizations might be using Python for data analysis and network visualization.
  • Have any of you used Visidata? I find the entire concept very intriguing and have been exploring it a little. I can see it being a very powerful tool that might be quicker for summary data analysis than trying to use Excel.
  • D-Tale is another new tool that provides a way to visualize and analyze pandas data. The basic concept is that it is a lightweight web client on top of a pandas DataFrame. Interesting!
  • Talk Python to Me episode 252 is a nice complement to my Escaping Excel Hell episode . It’s worth checking out if you have not heard it already.
  • I am a fan of Altair but have also been experimenting with Plotly Express recently. I am quite impressed with how intuitive it is. I think the Plotly team has developed a nice API that is very pythonic and familiar. It is very reminiscent of Seaborn. If you have only worked with Plotly in the past, you might want to check it out.

Final Words

I hope you enjoy the newsletter. If you have topics you would like to see covered in the blog or in the newsletter. Let me know by replying to this email. If you are new to the newsletter, the archive of past issues is available.