Book Image

Python Business Intelligence Cookbook

Book Image

Python Business Intelligence Cookbook

Overview of this book

The amount of data produced by businesses and devices is going nowhere but up. In this scenario, the major advantage of Python is that it's a general-purpose language and gives you a lot of flexibility in data structures. Python is an excellent tool for more specialized analysis tasks, and is powered with related libraries to process data streams, to visualize datasets, and to carry out scientific calculations. Using Python for business intelligence (BI) can help you solve tricky problems in one go. Rather than spending day after day scouring Internet forums for “how-to” information, here you’ll find more than 60 recipes that take you through the entire process of creating actionable intelligence from your raw data, no matter what shape or form it’s in. Within the first 30 minutes of opening this book, you’ll learn how to use the latest in Python and NoSQL databases to glean insights from data just waiting to be exploited. We’ll begin with a quick-fire introduction to Python for BI and show you what problems Python solves. From there, we move on to working with a predefined data set to extract data as per business requirements, using the Pandas library and MongoDB as our storage engine. Next, we will analyze data and perform transformations for BI with Python. Through this, you will gather insightful data that will help you make informed decisions for your business. The final part of the book will show you the most important task of BI—visualizing data by building stunning dashboards using Matplotlib, PyTables, and iPython Notebook.
Table of Contents (12 chapters)
Python Business Intelligence Cookbook
Credits
About the Author
About the Reviewer
www.PacktPub.com
Preface
Index

Creating reports in Excel directly from a Pandas DataFrame


In this recipe, you'll learn how to create an Excel report directly from a Pandas DataFrame using the to_excel() function. We will be writing all the code in IPython Notebook.

How to do it…

  1. First, import the Python libraries that you need:

    from pymongo import MongoClient
    import pandas as pd
    from time import strftime
  2. Next, create a connection to MongoDB and specify the accidents collection:

    client = MongoClient('localhost', 27017)
    db = client.pythonbicookbook
    collection = db.accidents
  3. Once you've created the connection, run a query to retrieve the first 1000 records where an accident happened on a Friday:

    data = collection.find({"Day_of_Week": 6}).limit(1000)
  4. Next, create a new DataFrame from the MongoDB query:

    df = pd.DataFrame(list(data))
  5. Show the first five rows to ensure that you have the right data:

    df.head()
  6. Now delete the _id column as we don't need it, and we can't write to the Excel file with it. The reason we cannot write it is because...