experimenta es com grandes volumes de dados usando
play

Experimentaes com grandes volumes de dados usando Notebooks - PDF document

qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Experimentaes com grandes volumes de dados usando Notebooks Gilmar Souza - Data & Analytics Principal 1 of 17 5/9/18, 11:57 AM qconsp18-gilmar


  1. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Experimentações com grandes volumes de dados usando Notebooks Gilmar Souza - Data & Analytics Principal 1 of 17 5/9/18, 11:57 AM

  2. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Notebooks 2 of 17 5/9/18, 11:57 AM

  3. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... 3 of 17 5/9/18, 11:57 AM

  4. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... source: http://jupyter.readthedocs.io/en/latest/architecture/how_jupyter_ipython_work.html (http://jupyter.readthedocs.io/en/latest /architecture/how_jupyter_ipython_work.html) 4 of 17 5/9/18, 11:57 AM

  5. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Por que Notebooks? REPL: Repeat-Eval-Print Loop 5 of 17 5/9/18, 11:57 AM

  6. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... In [18]: import time def paused_print(words_csv, delay): words = words_csv.split(',') for word in words: print(word) time.sleep(delay) paused_print('read,eval,print', 1) read eval print In [2]: #source: https://anaconda.org/jbednar/plotting_pitfalls/notebook import numpy as np np.random.seed(42) import holoviews as hv hv.notebook_extension('matplotlib') % opts Points [color_index=2] (cmap="bwr" edgecolors='k' s=50 alpha=1.0) % opts Scatter3D [color_index=3 fig_size=250] (cmap='bwr' edgecolor='k' s=50 alpha=1.0 % opts Image (cmap="gray_r") { + axiswise} % opts RGB [bgcolor="black" show_grid= False ] import holoviews.plotting.mpl holoviews.plotting.mpl.MPLPlot.fig_alpha = 0 holoviews.plotting.mpl.ElementPlot.bgcolor = 'white' from holoviews.operation.datashader import datashade from colorcet import fire datashade.cmap=fire[50:] Multimedia In [3]: def blues_reds(offset=0.5,pts=300): blues = (np.random.normal( offset,size=pts), np.random.normal( offset,size=pts reds = (np.random.normal( - offset,size=pts), np.random.normal( - offset,size=pts return hv.Points(blues, vdims=['c']), hv.Points(reds, vdims=['c']) blues,reds = blues_reds() blues + reds + reds * blues + blues * reds Out[3]: source: https://anaconda.org/jbednar/plotting_pitfalls/notebook (https://anaconda.org/jbednar/plotting_pitfalls/notebook) 6 of 17 5/9/18, 11:57 AM

  7. − 1 = ( x − 1) ( x + 1) x 2 qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... In [4]: hmap = hv.HoloMap({0:blues,0.000001:reds,1:blues,2:reds}, kdims=['level']) hv.Scatter3D(hmap.table(), kdims=['x','y','level'], vdims=['c']) Out[4]: In [19]: from ipywidgets import interact from sympy import Symbol, Eq, factor from sympy import init_printing init_printing() x = Symbol('x') def factorit (n): return Eq(x ** n - 1, factor(x ** n - 1)) interact(factorit, n=(2,20)); n 2 source: https://github.com/jupyter-widgets/ipywidgets/blob/766cad54a47c07520e9d695534c4664c3391e7ec/docs/source /examples/Factoring.ipynb (https://github.com/jupyter-widgets/ipywidgets/blob/766cad54a47c07520e9d695534c4664c3391e7ec /docs/source/examples/Factoring.ipynb) 7 of 17 5/9/18, 11:57 AM

  8. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... In [22]: import matplotlib.pyplot as plt from scipy.optimize import curve_fit as cf from ipywidgets import interactive, fixed, interact_manual from IPython.display import display N_samples = 25 x=np.linspace( - 2,2,N_samples) def f(x,a,mu,sigma): r=a * np.exp( - (x - mu) ** 2 / (2 * sigma ** 2)) return (r) def func(amplitude,ideal_mu,ideal_sigma,noise_sd,noise_mean): r=amplitude * np.exp( - (x - ideal_mu) ** 2 / (2 * ideal_sigma ** 2)) plt.figure(figsize=(8,5)) plt.plot(x,r,c='k',lw=3) r= r + np.random.normal(loc=noise_mean,scale=noise_sd,size=N_samples) plt.scatter(x,r,edgecolors='k',c='yellow',s=60) plt.grid( True ) plt.show() return (r) In [23]: y=interactive(func,amplitude=[1,2,3,4,5],ideal_mu=( - 5,5,0.5), ideal_sigma=(0,2,0.2), noise_sd=(0,1,0.1),noise_mean=( - 1,1,0.2)) display(y) amplitude 1 ideal_mu 0.00 ideal_sigma 1.60 noise_sd 0.40 noise_mean -1.00 source: https://towardsdatascience.com/interactive-machine-learning-make-python-lively-again-a96aec7e1627 (https://towardsdatascience.com/interactive-machine-learning-make-python-lively-again-a96aec7e1627) 8 of 17 5/9/18, 11:57 AM

  9. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Reproducibilidade e Colaboração 9 of 17 5/9/18, 11:57 AM

  10. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Dados tem Inércia 10 of 17 5/9/18, 11:57 AM

  11. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... 11 of 17 5/9/18, 11:57 AM

  12. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... In [8]: import io import base64 from IPython.display import HTML video = io.open('img/emr_zeppelin.mp4', 'r+b').read() encoded = base64.b64encode(video) video_data='''<video alt="test" controls> <source src="data:video/mp4;base64,{0}" type="video/mp4" /> </video>'''.format(encoded.decode('ascii')) 12 of 17 5/9/18, 11:57 AM

  13. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... In [9]: HTML(data=video_data) Out[9]: 13 of 17 5/9/18, 11:57 AM

  14. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... O que for preciso 14 of 17 5/9/18, 11:57 AM

  15. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... 15 of 17 5/9/18, 11:57 AM

  16. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Perguntas 16 of 17 5/9/18, 11:57 AM

  17. qconsp18-gilmar http://localhost:8888/notebooks/qconsp18-gilmar.... Gilmar Souza gilmar.souza@ifood.com.br (mailto:gilmar.souza@ifood.com.br) https://github.com/gilmar/qconsp18 (https://github.com/gilmar/qconsp18) http://gilmar.me (http://gilmar.me) 17 of 17 5/9/18, 11:57 AM

Recommend


More recommend