Monday, September 19, 2016

Summary of Findings: Trend Analysis (4 out of 5 Stars)

Note: This post represents the synthesis of the thoughts, procedures and experiences of others as represented in the articles read in advance (see previous posts) and the discussion among the students and instructor during the Advanced Analytic Techniques class at Mercyhurst University in September 2016 regarding Trend Analysis as an Analytic Technique specifically. This technique was evaluated based on its overall validity, simplicity, flexibility and its ability to effectively use on structured data.

Description:

Trend analysis is an analytical method that takes historical data and measures geographical and temporal changes among others. Based on these changes, a forecast can be made about potential future changes. Additionally, trend analysis can be used to isolate past events for further interpretation.

Strengths:

  • Increases forecasting accuracy
  • Data is flexible to use and manipulate pending the size of a given source’s data set
  • Supported by hard data but can also be adapted to forecast unstructured data to a degree
  • Presents visual perspectives that are easy to interpret
  • Simple to conduct analysis

Weaknesses:

  • Does not work as well with unstructured data
  • Susceptible to data tampering (cherry picking)
  • Dependent on the reliability of a data sources method of reporting
  • Does not always capture variability in data, and may propose a consistent forecast
  • Usually requires software to complete effectively

How-To:

  1. Collect measurable data that can be aggregated and utilized for analysis
  2. Input that data into a database tool that can be imported to other analytical programs
  3. Clean data so that it is correctly categorized and labeled to minimize potential errors
  4. Import data into analytical tool to oversee long-term and short-term changes
  5. Review data and set appropriate boundaries (time, location, etc.) to limit data for concise analysis
  6. Use forecasting analysis tools to predict future based on trend found within boundaries

Application of Technique:

Members of the class led a demonstration on how to create a trend analysis in the software Tableau. Data used consisted of Apple stock prices (open, close, adjusted close, high, and low) from the time period 1981-2016, and also pulled crime data from Chicago’s crime data portal on assaults in the Austin community area from 2002-2016. A trend of Apple’s stock price over time was created and a forecast was made to estimate future stock price changes over time. For the law enforcement data, observations were geocoded and used to create a heat map of incidents and how they changed in different areas over time.

For Further Information:

Investopedia Trend Analysis Definition: http://www.investopedia.com/terms/t/trendanalysis.asp

Trend Analysis Cliff Notes:

Trend Analysis for forecasting (+Tableau):

Google Trends:

Google Ngrams:

IACA - Types of Crime Analysis (Trend Analysis):

Historical Stock Data:
http://www.stockhistoricaldata.com/download

2 comments: