Interactive analysis of data using Pandas, Matplotlib in an iPython notebook

Transcript

1. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Interactive analysis of data using Pandas, Matplotlib in an IPython Notebook. Lakshman Prasad (twitter.com/becomingGuru) The Fifth Elephant July 13, 2013, Bangalore
2. These slides are online: http://j.mp/fifthel-python
3. Analysis notebook viewable online: http://j.mp/fifthel-usesthis
4. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Intro Demo Case for Python for Data Sneakpeek into upcoming Summary
5. How many of you use Hadoop Clusters?
6. How many of you use Hadoop Clusters for data under 500 GB?
8. ”Big Data” term encompasses a spectrum of tools and technologies
9. IPython notebook
11. Interviews of 373 influential people, 72 categories, 5066 user-wares
12. on the Hardware they use ...
13. ... and the Software
14. Demo Time!
15. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Popular Headphones!
16. US social security Database: Popular names
17. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary DataFrame all Names years = range (1880 ,2012) frames = [ ] for year in years : frame = pd . r e a d c s v ( ’ / path /yob%d . t x t ’%year , names=[ ’name ’ , ’ sex ’ , ’ b i r t h s ’ ] ) frame [ ’ year ’ ] = year frames . append ( frame ) names = pd . concat ( frames , i g n o r e i n d e x=True ) t o t a l b i r t h s = names . p i v o t t a b l e ( ’ b i r t h s ’ , \ rows=’ year ’ , c o l s=’ sex ’ , aggfunc=sum) t o t a l b i r t h s . p l o t ()
18. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Births across the century
19. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Popularity changing across the century
20. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Plot popularity names p = names . p i v o t t a b l e ( ’ b i r t h s ’ , rows=’name ’ , aggfunc=sum) top names = names p . order ()[ −5:] names by year = names . p i v o t t a b l e ( ’ b i r t h s ’ , rows=’ year ’ , c o l s=’name ’ , aggfunc=sum) names by year [ top names ] . p l o t ( s u b p l o t s=True )
21. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Case for Python for Data
22. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Supports parallel processing in clusters
23. Packages on networking, web, IO...
24. Managing clusters of machines (unless you need so many) is a chore
25. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary XKCD.com/552
26. Already a prevelant wisdom
27. ”Don’t go where the puck is; go where the puck is going to be”
28. Newer, Faster, Better versions of the entire stack
29. Rapid pace of development
31. Dynamic D3 visualizations from within the notebook.
32. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary In Summary • Moores Law curve stands at pretty good levels today.
33. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary In Summary • Moores Law curve stands at pretty good levels today. • Beyond a certain point, what you need is not necessarily more data, but right data.
34. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary In Summary • Moores Law curve stands at pretty good levels today. • Beyond a certain point, what you need is not necessarily more data, but right data. • Python data tools support clustering. Utilise it when you need it.
35. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary In Summary • Moores Law curve stands at pretty good levels today. • Beyond a certain point, what you need is not necessarily more data, but right data. • Python data tools support clustering. Utilise it when you need it. • Tools in Python: Pandas, Numpy, Matplotlib are potent toolkit for your data problems
36. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary In Summary • Moores Law curve stands at pretty good levels today. • Beyond a certain point, what you need is not necessarily more data, but right data. • Python data tools support clustering. Utilise it when you need it. • Tools in Python: Pandas, Numpy, Matplotlib are potent toolkit for your data problems • Python tools are improving themselves at a rapid pace
37. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary In Summary • Moores Law curve stands at pretty good levels today. • Beyond a certain point, what you need is not necessarily more data, but right data. • Python data tools support clustering. Utilise it when you need it. • Tools in Python: Pandas, Numpy, Matplotlib are potent toolkit for your data problems • Python tools are improving themselves at a rapid pace • Python data toolkit is very good at interoperability with other technologies - particularly, R
38. Either selling you something, or ...
39. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Dilbert gets it right!
40. Intro Demo Case for Python for Data Sneakpeek into upcoming Summary Image Attributions http ://www. f l i c k r . com/ photos / h i n k e l s t o n e /2765597758 http ://www. f l i c k r . com/ photos /25521149@N03/2491915744 http ://www. f l i c k r . com/ photos /50764322@N04/8378057370 http ://www. f l i c k r . com/ photos /micahdowty /3866958782 http ://www. f l i c k r . com/ photos /81512202@N05/7468442446 http ://www. f l i c k r . com/ photos / marlenek /2051394370 http ://www. f l i c k r . com/ photos / t e j e d o r o d e l u z /3157690060/ http ://www. f l i c k r . com/ photos /23820645@N05/4287681570/ http ://www. f l i c k r . com/ photos / a i d a n j o n e s /3575000735/ http ://www. f l i c k r . com/ photos /go /253819/ http ://www. f l i c k r . com/ photos / o r i n r o b e r t j o h n /114430223/
41. hello@lakshmanprasad.com