Status Updates From Python for Data Analysis

Data Aggregation, Group Operations, Time Series, and Financial and Economic Data Implications

Data Wrangling, clean, transform, merge, reshape

Currently using the 3rd generation of this book

DNF, alas.

(finshed chap.6)

158

Пройшов бібліотеку "Numpy". Нічого цікавого і малоймовірно, що знадобиться, але краще це знати і знати де шукати

After the first four chapters, skipped around to content I was most interested. Wes McKinney is a very good communicator of technology - helps that he created a good amount that he writes about.

Anyway, returned the library book well over a year ago.

Tons of examples, although it reads like a reference book.

Tác giả quyển sách cũng chính là người đóng góp chính cho thư viện pandas của python.

Slicing

Dry as a bone.

Next Chapter: 9.2 Plotting with panads and seaborn

Next: Chapter 9

Done Ch7, ~115 code cells, TODOs left. pandas changed Series.str behavior a bit since then, which is slightly frustrating: I need string&regex functionality for Pokémon tidying.

Without time structure & Pomodoro, doing anything is slow, mainly due to frequent short distractions. Meditation becomes more & more important.

Ch6, ~66 code cells, some TODOs left.

It does feel less like a “tutorial” & more of a showcase/gallery. & even then it feels lazy: he overviews a bunch of various features, but then lists the rest in tables, of which there are many.

~170 code cells, many TODOs left. My brain is burning, gonna go to bed early today. I don't know how I feel about the book yet, IMO not enough conceptual discussion, too much of “do this command, now do this command”.

Hopefully will start meditating/diet/jogging tomorrow.

Neiborinos watching “Inception” lol.

~145 code cells for Ch4, tho many TODOs left. Probably most of the Ch is about dtypes, dimensions, indexing (inc Boolean & fancy). Vectorization is told to be awesome, but the Secs on actual ufunc application are quite short. The book is primarily about pandas, after all, not linalg/stats/ML.

~121 code cells (but some TODOs left). I don't think the book is sufficiently newbie-friendly, there's some jargon here & there + concepts that may turn out to be quite tricky, but breezed thru too quickly. But he does talk about a lot cool stuff: comprehensions, generators, binary set operators, lambdas.

That was a wild ride across Python features, my Jupyter Notebook of Ch2 has 69 code cells. (Ecuadorians are partying in the living room, I listen to Brace Belden “War is Heck” episode)

Ch3 will be about Python containers, files, functions. A lot of intersection with “Automate the Boring Stuff…”

Book uses Python 3.6 & Anaconda (good). Dummy calls Emacs and Vim primitive, LOL.

Btw, Wes McKinney is the creator of pandas. In Preface he acknowledges (In Memoriam) late John D. Hunter, who created matplotlib. He died shortly after Wes completed 1st edition manuscript in 2012.

ToC indicates the book briefly introduces Python & IPython, then moves quickly to basic NumPy in Ch4 & immediately to pandas in Ch5. I think for now I should be interested in up to Ch10 for Pokémon stuff (but Ch12 has important pandas stuff like method chaining). Ch11 is Time Series, Ch13 is statsmodels & scikit-learn, Ch14 is example data for practice, appendices are advanced NumPy & IPython.

Next: DataFrame

Next: Introspection

Next chapter: Python Language Basics, IPython, and Jupyter Notebooks

Going to finish this and update my blog and resume real soon now, still need a new job.

Four more chapters, I should get at least one done tomorrow before my HackerX event.

I wanted to be done this book by now, but I typed most of the code out by hand while reading it... I really need a new job, tomorrow I network, but tonight I hope to get another chapter at least read.

"CHAPTER 6 Data Loading, Storage, and File Formats" is read