|
@DATAMAX
|
|
Applied Statistics for Software Managers |
|
Published by Prentice Hall
PTR, June 2002. Chinese translation published by Tsinghua University Press, June 2006 Click here to read reviews and see inside the book on amazon.com. |
You’ve implemented a
measurement program and have collected some software metrics data. Great, but
do you know how to make the most of this valuable asset? Categorical variables
such as language, development platform, application type and tool use can be
important factors in explaining the cost, duration and productivity of your company’s
software projects. However, analyzing a database containing many non-numerical
variables is not a straightforward task.
Statistics, like software
development, is as much an art as it is a science. Choosing the appropriate
statistical methods, selecting the variables to use, creating new variables,
removing outliers, picking the best model, detecting confounded variables,
choosing baseline categorical variables, and handling influential observations
all require that you make many decisions during the data analysis process.
Decisions for which there are often no clear rules. What should you do? Read
on.
Using real software project
data, this book leads you through all the steps necessary to extract the most
value from your data. In Chapter 1, I describe my methodology for analyzing
software project data. You do not need to understand statistics to follow the
methodology. I simply explain what to do, why I do it, how to do it and what to
watch out for at each step.
Common problems that occur
when analyzing real data are thoroughly covered in four case studies of
gradually increasingly complexity. Each case study is based around a business
issue of interest to software managers. In Chapter 2, you will learn how to
determine which variables explain the differences in software development
productivity. In Chapter 3, you will look at factors which influence
time-to-market. In Chapter 4, you will learn how to develop and measure the
accuracy of cost estimation models. In Chapter 5, you will study the cost
drivers of software maintenance, with an emphasis on presenting results.
Finally, in Chapter 6 you will learn what you need to know about descriptive
statistics, statistical tests, correlation analysis, regression analysis and
analysis of variance.
Outstanding features of the book:
- Four case studies: Software Development Productivity, Time to Market, Cost Estimation, and Software Maintenance Cost Drivers, provide examples of statistical methods using real software project data. This data has never been published and the book will make it publicly available.
- In the first chapter a Recipe for analyzing data is provided. This recipe can be followed without understanding the mathematics of the statistical methods employed. This will be very useful for software managers who need an answer fast.
- The details of the data analysis methodology I've developed have never been published.
- Everything is explained in simple terms, as if I were giving private tuition to the reader.
Market
Considerations:
I wrote this book for current
and future software managers. In particular, the unique combination of
statistics applied to software business issues should help every future
software engineer/manager understand why software measurement is useful and
what to do with the data.
The book should be of interest to:
- software managers because it is the only book they will need to understand how to analyze their software project data.
- researchers because of the real software project data it contains, and the data analysis methodology.
- professors (and students) of software measurement courses
- professors (and students) of general statistics courses who wish to make their course more up-to-date by using examples from the software industry.
The book could be used as the basis for a corporate training program, and in the software engineering and information systems curricula of universities. Additionally, it could be used in statistics courses taught to computer scientists as it contains examples of more interest to them.
Prerequisites: Anyone who wants to analyze data will need to know how to use a statistical software tool. As far as mathematics go, a basic knowledge of algebra is sufficient.