This shows you the differences between two versions of the page.
Next revision | Previous revision | ||
cs501r_f2016:lab1 [2016/08/29 16:27] admin created |
cs501r_f2016:lab1 [2021/06/30 23:42] (current) |
||
---|---|---|---|
Line 16: | Line 16: | ||
{{:cs401r_w2016:lab1.png?nolink|}} | {{:cs401r_w2016:lab1.png?nolink|}} | ||
- | **Part 2:** Your notebook should use the pandas library to read in the Rossman store sales data (a CSV dataset) and plot the sales of store #1. Your plot should look something like this: | + | **Part 2:** You must play with the Tensorflow playground neural network, and figure out how to create a classifier that successfully classifies the "spiral" dataset. |
- | {{:cs401r_w2016:lab1_storesales.png?direct&700|}} | + | [[http://playground.tensorflow.org/|Tensorflow playground]] |
- | + | ||
- | Done correctly, this should only take a few lines of code. | + | |
---- | ---- | ||
Line 28: | Line 26: | ||
* 20% Successfully turned in a notebook with working code | * 20% Successfully turned in a notebook with working code | ||
- | * 20% Random image with 50 random elements | + | * 35% Random image with 50 random elements |
- | * 20% Correctly used pandas to load store sales data | + | * 35% Image indicating tensorflow success |
- | * 30% Some sort of plot of sales data (only for store #1!) | + | |
* 10% Tidy and legible figures, including labeled axes where appropriate | * 10% Tidy and legible figures, including labeled axes where appropriate | ||
Line 37: | Line 34: | ||
Throughout this class, we will be using a combination of ipython | Throughout this class, we will be using a combination of ipython | ||
- | notebooks and the anaconda python distribution. For this lab, you | + | notebooks, Tensorflow and the anaconda python distribution. For this lab, you |
must install anaconda, and write a simple python program (using | must install anaconda, and write a simple python program (using | ||
- | ipython notebooks). As described above, the notebook should do two things: | + | ipython notebooks). |
- | 1) generate simple random images, and 2) plot some data using pandas. | + | |
+ | As described above, the notebook should do two things: | ||
+ | 1) generate simple random images, and 2) display an image that you generate using the Tensorflow playground. | ||
For part 1, you can generate any sort of random image that you want -- consider | For part 1, you can generate any sort of random image that you want -- consider | ||
Line 50: | Line 49: | ||
random images with moderate complexity. | random images with moderate complexity. | ||
- | In preparation for future labs, we strongly encourage you to use the | + | For part 2, you should visit the Tensorflow playground (see link above), and play with different settings. Most of it will be unfamiliar, but don't worry -- you can't break it! |
- | [[http://cairographics.org/|cairo]] package as part of your image generator. | + | |
- | + | ||
- | For part 2, the data you should use is downloadable here: | + | |
- | [[http://hatch.cs.byu.edu/courses/stat_ml/store_train.csv|Rossman store sales data]] | + | Once you have a working classifier, take a screenshot. Then use your ipython notebook to display that image in-line. |
---- | ---- | ||
Line 62: | Line 58: | ||
http://docs.continuum.io/anaconda/install | http://docs.continuum.io/anaconda/install | ||
- | To generate images, check out PIL and cairo: | + | To generate images, check out PIL. |
- | + | ||
- | ''conda install cairo'' | + | |
To generate random numbers, check out the [[http://docs.scipy.org/doc/numpy-1.10.0/reference/routines.random.html|numpy.random]] module. | To generate random numbers, check out the [[http://docs.scipy.org/doc/numpy-1.10.0/reference/routines.random.html|numpy.random]] module. | ||
Line 116: | Line 110: | ||
nbimage( data ) | nbimage( data ) | ||
</code> | </code> | ||
- | |||
- | ---- | ||
- | ====Using Pandas:==== | ||
- | |||
- | For the second part of this lab, you will need to understand the ''pandas'' python package, just a little bit. For this lab, you only need to know how to select some data from a CSV file. | ||
- | |||
- | You should read through this tutorial and play with it. | ||
- | |||
- | [[http://synesthesiam.com/posts/an-introduction-to-pandas.html|Tutorial on using Pandas]] | ||
- | |||
- | For this lab, you need select the data for store #1 and plot it. | ||
- | |||
- | An important part of generating visualizations is conveying information cleanly and accurately. You should therefore label all axes, and in particular, the x-axis should be labeled using dates (See the example image). This involves a bit of python trickery, but check out some helpful functions in the hints below. | ||
---- | ---- | ||
Line 138: | Line 119: | ||
import matplotlib.pyplot as plt | import matplotlib.pyplot as plt | ||
- | plt.plot_date | ||
- | |||
- | pandas.to_datetime | ||
plt.legend | plt.legend | ||
Line 148: | Line 126: | ||
plt.tight_layout | plt.tight_layout | ||
+ | </code> | ||
+ | |||
+ | Also note that to get plots to show up inline, you may have to add the magic incantation **in the first cell**: | ||
+ | |||
+ | <code python> | ||
+ | |||
+ | %matplotlib inline | ||
+ | |||
+ | import matplotlib | ||
+ | import numpy as np | ||
+ | import matplotlib.pyplot as plt | ||
+ | |||
+ | </code> | ||
+ | |||
+ | Here is a full setup for cairo in linux. | ||
+ | <code> | ||
+ | sudo apt-get install libcairo2-dev | ||
+ | git clone https://github.com/pygobject/pycairo.git | ||
+ | cd pycairo/ | ||
+ | python setup.py build | ||
+ | python setup.py install | ||
+ | </code> | ||
+ | |||
+ | Here is a setup for cairo in mac. | ||
+ | <code> | ||
+ | brew install cairo --use-clang | ||
+ | brew install py2cairo | ||
</code> | </code> | ||