1. Learn
    2. /
  2. Courses
    3. /
  3. Manipulating Time Series Data in Python
  • 1

    Working with Time Series in Pandas

    Free

    This chapter lays the foundations to leverage the powerful time series functionality made available by how Pandas represents dates, in particular by the DateTimeIndex. You will learn how to create and manipulate date information and time series, and how to do calculations with time-aware DataFrames to shift your data in time or create period-specific returns.

  • 2

    Basic Time Series Metrics & Resampling

    This chapter dives deeper into the essential time series functionality made available through the pandas DataTimeIndex. It introduces resampling and how to compare different time series by normalizing their start points.

  • 3

    Window Functions: Rolling & Expanding Metrics

    This chapter will show you how to use window function to calculate time series metrics for both rolling and expanding windows.

  • 4

    Putting it all together: Building a value-weighted index

    This chapter combines the previous concepts by teaching you how to create a value-weighted index. This index uses market-cap data contained in the stock exchange listings to calculate weights and 2016 stock price information. Index performance is then compared against benchmarks to evaluate the performance of the index you created.

Initializing

Exercise

Random walk II

In the last video, you have also seen how to create a random walk of returns by sampling from actual returns, and how to use this random sample to create a random stock price path.

In this exercise, you'll build a random walk using historical returns from Facebook's stock price since IPO through the end of May 31, 2017. Then you'll simulate an alternative random price path in the next exercise.

Instructions

100 XP

We have already imported pandas as pd, choice and seed from numpy.random, seaborn as sns, and matplotlib.pyplot as plt. We have also imported the FB stock price series since IPO in May 2012 as the variable fb. Inspect this using .head().

  • Set seed to 42.
  • Apply .pct_change() to generate daily Facebook returns, drop missing values, and assign to daily_returns.
  • Create a variable n_obs that contains the .count() of Facebook daily_returns.
  • Use choice() to randomly select n_obs samples from daily_returns, and assign to random_walk.
  • Convert random_walk to a pd.Series and reassign it to itself.
  • Use sns.distplot() to plot the distribution of random_walk.