NYC Open Data

Name: Python Workshop II: Pandas for data analysis(no python basic)
Start: 2013-09-16T23:15:00Z
Location: On Deck Capital

Python Workshop II: Pandas for data analysis(no python basic)

Sep 16, 2013 · 7:15 PM
On Deck Capital

This is a enhanced session of Aug 26th "Python Workshop I: Pandas for data analysis(with python basic)". You can find useful information from comments at previous event.
http://www.meetup.com/NYC-Open-Data/events/128896482/

Many Thanks to On Deck Capital for hosting us and John Downs for sharing his experience with Pandas.

------------------------------------------------------
Speaker Bio:

John Downs is a software engineer at On Deck. He has been programming in Python for several years. John also has interests in Machine Learning and Data Mining.

Preparation:

Bring your favorite editor and python shell. Install pandas with 'pip install pandas'.

Agenda:

7-8: We'll dive into using Pandas to do some analysis and visualization directly.
8-9: We'll break into groups to give everyone a chance to apply the material from the talk.

Join or login to comment.

Hit enter to add your reply

A former member

Hey! I know it's very very very late, but here is the notebook from this talk:

http://nbviewer.ipython.org/urls/raw.github.com/jdowns/Presentations/master/pandas_ii/nyc_open_data_pandas_ii.ipynb

:)

1 · November 18, 2013
Shangxuan Vivian Z.

Hi attendants, I felt bad for the fact that a few of you left early. Could you be kindly share your reason? Thanks!

0 · September 16, 2013
- View all 3 replies
- André P.
  
  I sat by the door and saw their faces. They didn't seem to say "the heck with it, this is completely useless", but rather, "this is good, even if a bit puzzling; I have to get more into it". I suppose their reasons for leaving early were outside: trains to catch, dogs to walk, kids crying hungry, wives hating their guts for staying away from home in order to improve their chances to bring a paycheck that will buy their dresses, or other reasons that wouldn't cross anyone's mind. The best way to know is to ask, which you just did.
  
  0 · September 16, 2013
- A former member
  
  The reason that I left early is obvious: the talk was too advanced for my rudimentary Python skills.
  
  My plan was to expose myself to a professional caliber program in hopes that it would be within striking distance of my rudimentary skills, and I would thereby be lifted up; such strategy was wrong ... "the best laid plans of mice and men" ...
  
  I did indeed derive one benefit: I now know that there's alot more to Python than the elementary sites - Udacity, eg - would have the tyro believe.
  
  If you can recommend a more advanced online learning site, one that will raise my skills closer to what is required by a professional program, then please apprise me of such
  
  0 · September 19, 2013
- Karen B.
  
  You may want to jump in on this course https://www.edx.org/co.... It's only in wk 2. Starts at the beginning but gets complex quickly.
  
  0 · October 28, 2013
André P.

John: Any chance you could post a link to the notebook and/or Python code from the talk?

0 · September 18, 2013
Xu F.

Wish that the code was shared online during the tutorial. I find it easier to follow that way, just like other R intro workshops.

0 · September 17, 2013
Dev

Performance metrics:
http://wesmckinney.com/blog/?p=414
http://pandas.pydata.org/pandas-docs/vbench/

0 · September 16, 2013
André P.

Vivian, John:

You know that I wasn't happy with the first Pandas talk, which turned out to be most entirely an introduction to Python. For your next talk(s), I would like to see the following items:
- no Python; those who know nothing about Python belong in a different meetup before attending this one (and nothing can stop you from presenting such a Python-intro meetup just before it, if you think there is sufficient demand)
- code to download a file
- code to open a local file in comma-delimited, tab-delimited, JSON, Excel and SPSS formats
- code to save a file locally in the above formats, where available
- a short discussion of NumPy arrays and how to create them from Python objects, with iterators, and with random number generators
- basic Pandas types: Series, TimeSeries, DataFrame, Panel; underlying Pandas/NumPy element types for int, float, string, datetime; handling of missing values
- various methods of selecting subsets of rows and/or columns

(continued)

2 · September 15, 2013
- View all 7 replies
- André P.
  
  - vector-like operations; "broadcasting" if vector operands are not the same size (different from R? how?)
  - data manipulation: inserting and deleting columns, truncating, concatenating series and data frames, grouping, merging, joining, matrix and matrix-vector multiplication
  - basic Matplotlib charts: barchart (simple, grouped, stacked), piechart, boxplot, line chart, scatterplot, bubbleplot
  - annotating charts: adding titles, legends, text labels, lines, curves, arrows (thin and thick), dimension lines
  - 3D plots (optional): line chart, scatterplot, wireframe
  - maps (optional; this could be the subject of another meetup, not necessarily Python-based)
  - linear algebra: solving a linear system (general, symmetric), least squares
  - statistical procedures in Pandas and SciPy: descriptive summaries, interval estimation, hypothesis tests (Z, t, chi^2, ANOVA), regression
  - (time-)series operations: term-to-term differences, cumulative sums, autocorrelations, forecasting, etc.
  
  1 · September 15, 2013
- André P.
  
  Because this talk will introduce several packages new to most attendees, I strongly suggest that all packages be imported straight (with no "from" or "as" clauses) and all names be fully qualified, so it's clear which name comes from which package. Customary usage, like "from pandas import *" and "import matplotlib.pyplot as plt", should be presented at the end.
  
  The IPython notebook should be split into a Python part and a Pandas part, and the Python part relegated to the Python talk. You might also post the code as a Python file, for those without IPython.
  
  I realize this is a lot of material. Perhaps it might make sense to cover only the data manipulation and graphics now, and leave the number crunching for the Pandas II meetup.
  
  André
  
  1 · September 15, 2013
- Shangxuan Vivian Z.
  
  We will do the best! Thank you for these suggestions!
  
  0 · September 16, 2013
- Shangxuan Vivian Z.
  
  Andre, thank you for taking so much time to write a detailed expecting! I really mean it.
  
  0 · September 16, 2013
- Shangxuan Vivian Z.
  
  btw, it was my fault to ask John to cover basic python. There was one attendant who didn't know any python and wanted to come. So we did best to take him as guest.
  
  0 · September 16, 2013
- Dev
  
  I second the 'no python intro' rule. With only 2 hours, focus should be on learning pandas, not python.
  
  0 · September 16, 2013
- A former member
  
  I agree as well, focus should be on learning pandas
  
  0 · September 16, 2013
Angus Vincent W.

Unable to make it tonight, will there be another one.

0 · September 16, 2013
A former member

Looking forward to it!

0 · July 10, 2013
Paul A.

I am currently wait listed for the first session, so if I make it to that one I can free up my slot for another person to this event.

0 · July 10, 2013

60 went

Shangxuan Vivian Z.

Organizer,

Event Host
Emily
Rick B.
Daisy L.
in/aaronchall
Steven F.
Adam P.
Maria M.
Essell
Simcha L.
Daniel E.
Whittle
Phil H.
Craig C.
Charles L.
Steven S.
Jorge
Shubin L.
Dev
Kevin
Elliot N.
Ken A.
Saumil A.
Margo
srikrishna
Leo D.
Shreyas S.
David E.
Allister
Ling
Vivian C.
Kannan
Jules R.
Dinesh
Nitin k.
Viktor S.
Hans V.
Marco C.
Inwha K.
Xu F.
André P.
Carlos M.
A former member
A former member
A former member
A former member
A former member
A former member
A former member
A former member

NYC Open Data

New York, NY

Founded May 23, 2013

About us…

Help support your Meetup

Chip in

Organizers:

, , , ,

Contact

We're about:

· · · · · · · · · · · ·

Our Sponsors

NYC Data Science Academy

use"nycopen100ff" coupon to take classes on www.nycdatascience.com
Supstat

Supstat shares its expertise in data mining and visualization.

People in this
Meetup are also in:

#BetaNYC, NYC's open data, open gov, & civic tech community

4,257 Brigade Members
ny-scala

3,255 Scala Types
Metis: New York Data Science

5,634 Data Scientists
New York Data Science Study Group

3,819 Data Science Padawans
IoT Central® | NYC (Internet of Things Events)

5,688 IoTers
Big Data Warehousing

4,500 Data Nerds

Python Workshop II: Pandas for data analysis(no python basic)

A former member

A former member

A former member

A former member

60 went

Organizer,

Event Host

You're all set!

We're about:

People in this Meetup are also in:

Sign up

Sign up

People in this
Meetup are also in: