Data Science with Jupyter: 2-in-1

This video will give you an overview about the course.

Jupyter is available as a web application for a wide variety of platforms. This video covers the details of the Jupyter user interface: what objects it works with and what actions can be taken by Jupyter.

• Look at the Jupyter user interface
• Perform actions with Jupyter

In this video, we will see different menu choices on the menu bar.

• Study the menu choices

In this video, we will see several examples taken from current industry focus and apply them in Jupyter to ensure its utility. This video will explain European call option valuation and Monte Carlo pricing. We will also look at gambling for betting analysis.

• Implement Monte Carlo pricing
• Determine probability of series in coin flip

This video is all about real time examples such as insurance and consumer product. We are using R to see pricing for non-life products and marketing effectiveness.

• Implement non-life insurance pricing
• Look at effectiveness of different and campaigns for grape fruit juice

The predominant Jupyter hosting product currently is JupyterHub. It provides multi-user access to your notebooks.in this video, we will install JupyterHub. We will also see Jupyter hosting.

• Install JupyterHub
• Access JupyterHub installation

Optimizations cover a gamut of options running from language-specific issues to deploying your notebook in a highly available environment. Optimizations are script language dependent.

• Use timeit() to determine execution time
• Use profiler to give complete rundown of the execution

R also has tools available that will help pinpoint performance issues with your R coding such as Microbenchmark, modify a function used frequently. optimize name lookup, optimize data frame value extraction, R implementation and Change algorithm.

• Use microbenchmark to profile R script
• Look at caching the notebook concept

Securing a notebook can be accomplished by several methods such as: Manage authorization and Securing notebook content.

• Study issues with standard content in Jupyter
• Look at the techniques to overcome security issues

Python has several groups of processing functions that can tax computer system power. In this video, we will use NumPy function that is a package in Python providing multidimensional arrays and routines for array processing.

• Use NumPy function in jupyter

Pandas is an open source library of high-performance data analysis tools available in Python. We will see functions like read text files, read Excel files, read from SQL database and operate on data frames.

• Use pandas to read text files and excel files
• Use pandas to work with dataframe

SciPy is an open source library for mathematics, science and, engineering. We will see many areas that can be explores using SciPy such as Integration, Optimization, Interpolation, Fourier transforms and Linear algebra.

There are more functions built-in for working with data frames that we have used so far. If we were to take one of the data frames, we could use additional functions to help portray and work with the dataset.

• Use slicing to expand the panda data frame

Data frames automatically allow you to easily sort and filter the dataset involved, using existing functionality within the data frames themselves.

• Implement filtering based on certain criteria
• Implement sorting data frame by index

scikit-learn is a machine learning toolset built using Python. In scikit-learn, an estimator provides two functions, fit() and predict(), providing mechanisms to classify data points and predict classes of other data points, respectively.

• Implement prediction model using scikit learn

In this video, we will make a prediction using R. The functions are different for the different language, but the functionality is very close.

• Build prediction model using R

There is a Python package, Bokeh, that can be used to generate a figure in your notebook where the user can interact and change the figure. In this video, I am using the same data from the gridplot example to display an interactive Bokeh gridplot.

• Plot a graph using Plotly
• Create a human density map

In this video, we will gather one of the datasets and produce a histogram from the data.

• Plot 3D data using car dataset

Spark is a fast, general engine for large-scale data processing. The SparkContext initializes all of Spark and sets up any access that may be needed to Hadoop, if you are using that as well.

• Analyze number of lines in a file using spark

Spark exposes many SQL-like actions that can be taken upon a data frame. In our example we will start a Spark Session, uses the session to read a CSV formatted file, that contains a header record and finally displays initial rows.

• Use Spark SQL to determine product list

We will combine data frames, operate on it resulting set, import JSON data, and manipulate it with Spark.

• Populate the data frames and move it to spark

Spark can also access JSON data for manipulation. In this video, we will also see a pivot() that allows you to translate rows into columns while performing aggregation on some of the columns.

• Read and load the JSON in Spark
• Use pivot() for translation from row to column

To get a flavor of the resources available to R developers, we can look at the 2016 election data.

• Setting up R for Jupyter
• Display information about the data frame

In this video, we will look at voter registration versus actual voting using census data.

• Display information to visually check for accurate loading
• Display the characteristics of the regression line

In this video, we can look at trends in college admissions acceptance rates over the last few years.

• Create a vector of the average acceptance rates for colleges
• Convert the vector points into a time series

In this video, we will look at the airline arrival and departure times versus scheduled arrival and departure times.

• Build our model of the arrival time
• Use the testing set to make predictions
• Plot the predicted versus actual data

In this video, we will walk through the process of reading a CSV and adjusting the dataset to arrive at some conclusions about the data.

• Change the column names to be more readable
• Get rough statistics on the data
• Plot some dominant data points

In this video, we will use dplyr package against the baseball player statistics we used earlier.

• Convert a data frame to a dplyr table
• Filter rows in a data frame
• Add a column to a data frame

The tidyr package is available to clean up/tidy your dataset. In this video, we will rearrange our data to mix columns and rows with values.

• Use the standard example of stock price for a date
• Use the spread() function to separate out the values into multiple columns
• Reorganize the data by listing all prices for a stock per row

In this video, we will look to display glyphs at different points in a graph rather than the standard dot as the glyph should provide more visual information to the viewer.

• Display glyph data about the standard iris dataset
• Derive some information from the glyph data

You can publish a notebook/dashboard using markdown. Markdown involves adding annotations to cells in your notebook that are interpreted by Jupyter and converted into the more standard HTML representations that you see in other published materials. In this video, we will see the different kinds of markdowns.

• Create cell with the markdown type
• Generate table in HTML

Shiny is a web application framework for R. The Shiny server code set deals with accessing data, computing results, obtaining direction from the user and interacting with other server code set to changes results. In this video, we will learn how to create a shiny dashboard.

• Load the Shiny dashboard library
• Publish your dashboard

Using Node.js, developers have come up with a way to host your dashboard/notebook without Jupyter on jupyter-dashboard-server.

• Install conda
• Install the layout extension

In this video, we will use the Yelp data and use the dataset from round 9 of the challenge.

• Download the JSON file and upload it on Jupyter notebook
• Display the date and time of our system
• Create a reviews.csv

In this video, we will build a computed data frame with two columns and display the top-rated business dataset. Also, we will visualize the relationship between ratings and number of reviews for companies.

• Find the top-rated firms
• Build a model of reviews

Naive Bayes is an algorithm that uses probability to classify the data according to Bayes theorem for the strong independence of the features. Bayes theorem estimates the probability of an event based on prior conditions. So, overall, we will use a set of feature values to estimate a value assuming the same conditions hold true when those features have similar values. Also, we will implement naive Bayes using the R programming language.

• Install the package and load the library
• Measure the accuracy of the model
• Determine the accuracy of the model

Using the nearest neighbor, we will have an unclassified object and a set of objects that are classified. In this video, we will take the attributes of the unclassified object, compare against the known classifications in place and select the class that is closed to our unknown.

• Reorder the data in ascending order
• Split the data into a training set and a test set
• Use the same example and implement the nearest neighbor using Python

In this video, we will use decision trees to predict values. A decision tree has a logical flow where the user makes a decision based on attributes the tree down to a root level where a classification is provided.

• Load the libraries to use rpart
• Develop a model to predict mpg acceptability based on the other factors
• Perform the same analysis in Python

With a neural net, we will end up with a graphical model that provide the factors to apply to each input to arrive at our housing price. Also, we will use the random forest algorithm which attempts many random decision trees and provide the tree that works best within the parameters used to drive the model.

• Calculate our neuralnet model
• Include the packages of random forest in R