ArtsAutosBooksBusinessEducationEntertainmentFamilyFashionFoodGamesGenderHealthHolidaysHomeHubPagesPersonal FinancePetsPoliticsReligionSportsTechnologyTravel

Predictive Analytics using Excel 2010 (part 1) - Overview

Updated on August 7, 2012

What is Predictive Analytics?

In this series, I will be conducting predictive analysis using a hypothetical scenario where an individual wants to predict the future share price of a specific stock. Although this is a rather simple scenario, the overall process can be applied to complex situations where multiple variables/predictors are present.


Overview

Predictive analytics is a form of business intelligence that uses a variety of statistical techniques to predict future events or behaviors from current and historical data using predictors. Such predictors are combined, analyzed, and used for prediction with an acceptable level of reliability (or consistency). Data mining is a key component of predictive analytics which employs techniques such as regression (linear and logistic), time-series forecasting, weighted moving averages, and various descriptive statistics to create sophisticated predictive models.

To create such sophisticated models, data is collected, applicable statistical techniques are used to cluster various variables/predictors together, predictions are then made and validated, and the whole model is tweaked as new data comes to light.

Businesses often use predictive analytics to answer questions about customer behavior. For example, using such predictors as age, level of education, and buying habits, a business could determine whether a particular new customer will become a long-term customer or churn in the next few months. In short, businesses use predictive analytics to better understand their customers, products, and partners in order to identify future risks before they happen, find hidden opportunities, and optimize their current processes.

There are a number of different software solutions out there to conduct predictive analysis such as SAS, SPSS, and R; however, in this series I will be focusing on Excel 2010 as it’s relatively simple to use, and if used correctly, can be just as effective as many of the more powerful solutions.

Keep in mind that predictive analysis is just one part of business intelligence—

  1. Predictive Analytics – developing a forecast
  2. Dashboards – monitoring the forecast
  3. Analysis – analyzing deviations and variances from the forecast
  4. Reporting – reporting on the successes and failures of the initial forecast

Process

The following details the steps in Predictive Analysis—

  1. Define the problem or business need and the desired outcomes
  2. Determine the type of data needed and collect such data (if necessary)
  3. Prepare the data (transform, clean, etc.) for statistical analysis
  4. Conduct analysis and evaluate which analytical model meets the desired outcomes and addresses the business need
  5. Distribute the model to applicable business partners throughout the organization
  6. Continue to improve the reliability and accuracy of the model


Please Note: In this series, the focus is on data collection and statistical analysis


Next Steps

In Part 2, using a hypothetical scenario, the need for a predictive model is defined and data is collected.


Table of Contents

  1. Part 1 - Overview
  2. Part 2 - Needs Analysis and Data Collection

Comments

    0 of 8192 characters used
    Post Comment

    No comments yet.