Sunday 10:00–10:35 in Auditorium

Case Study in Travel Business - Time Series Analysis with Seasonal Data

Cheuk Ting Ho

Audience level:


For time series analysis, everyone's talking about ARIMA or Holt-Winters. But there's other models which could also break down a seasonal series into trend, seasonality and noise. We will use an open source Python library called Seasonal to analyse B2B worldwide travel data.


Times series analysis is an important part of data analysis for lots of businesses. It is very often for stakeholders to be interested in the performance of the business by analyzing measurements of profit, cost, number of sale, number of searches etc over time. In this talk, we will do a case study of showing how we estimate the impact public holidays made on the travel business. The method of analyzing the time series by seasonal breakdown will be explored and the work flow of solving the problem will be explained.

In the first half of the talk, an introduction about time series and its characteristic will be explained for audiences who is new to analysis on time series. The data we use will be from a business to business travel company. It has seasonality thought out the year, a weekly cycle and also a growing trend in business. As the company have clients around the world, data from different countries will shows different behaviors as well. Therefore, before we show the analysis, the complexity of the data will be explored. In the second half, we will introduce a open source Python library called Seasonal. Using this package, we will demonstrate how to break down the travel data and extract the fluctuation of the sale in different countries. By comparing the fluctuation and Google calendar, public holidays in different countries can be spotted and their impact on the business can be estimated.

This talk is for people who are interested in time series analysis and its application in business. Audiences with or without experience would also found this talk useful in giving them insights in how a business could benefit in making use of the data and doing a proper time series analysis.

Subscribe to Receive PyData Updates