randomosity

Data Ana­lysis with Open Source Tools by Phil­ipp K Jan­ert is a simply superb, solid and exhaust­ive syn­thesis of instruc­tion, work­shops and hands-on exer­cises designed for those ser­i­ous about con­duct­ing pro­fes­sional data ana­lysis. This is not a light­weight under­tak­ing. This is a ser­i­ous get-down-to-it and do-it-right kind of manual. The author (as has been men­tioned else­where) is pas­sion­ate about his sub­ject and it shows. He knows how to con­vey the most com­plex con­cepts in an approach­able and effect­ive way.

The title of the book sug­gests a very hands-on approach and sug­gests pos­sibly that it is more bout apply­ing known tech­niques util­ising spe­cific tools. The book deliv­ers on this through a series of ‘work­shops’ that are attached to each chapter. These sec­tions lead you through spe­cific applic­a­tion of spe­cific data ana­lysis tasks using par­tic­u­lar pack­ages (primar­ily Python and R) in a very use­ful man­ner. But don’t be deceived, the book is far more than these work­shops and to its credit provides extens­ive ground­ing in the the­ory and prin­ciples of data ana­lysis itself so as to ground you in the application.

Whether you are approach first attempts at data ana­lysis and feel unsure about the prac­tise, or you have been using simple tools such as excel to carry out your ana­lysis or are look­ing to hone your tech­niques by explor­ing the power of R is mov­ing to present­a­tion of your ana­lyt­ical find­ings, you will find this volume a superb choice. Spend­ing time work­ing through the work­shops will build a firm found­a­tion of cap­ab­il­ity to extend the the­or­et­ical and addi­tion­ally provides a superb ground­ing for approach­ing courses deal­ing more dir­ectly with the tools them­selves, such as the volumes on R that I have explored previously.

Jan­ert uses the pro­gres­sion of present­ing the data (to find the pat­terns) –> mod­el­ling the data (to explore) –> min­ing the data com­pu­ta­tion­ally (to under­stand the data) –> and finally apply­ing the data (to actu­ally use it in real world instances) through the book. He uses examples lib­er­ally to main­tain engage­ment and styl­ist­ic­ally asks the reader ques­tions and makes remind­ers con­stantly to keep you mov­ing through what I remind is pretty heavy mater­ial. This is not a book to try in one set­ting, nor how­ever, is it a ref­er­ence manual. It is really a course that should be approached over a suit­able length of time.

One of the ques­tions the author poses early on is ‘what’s with the math?’ and he assures you that if you do find this intim­id­at­ing its worth the time to famil­i­ar­ise and gain some com­fort with them as they are neces­sary should you really want to carry out effect­ive data ana­lysis. He’s right and the way the book is struc­tured you do need to take this advice on board. This will limit the book to people that are will­ing to com­mit if uncom­fort­able with the con­cepts. But as I men­tioned above this is a ser­i­ous book and the author seems to have made a com­mit­ment him­self to deliver the mater­ial and asks for a bit of recip­roc­a­tion. Now, I (pos­sibly less than fondly) recall much of this from the dis­tant haze of under­gradu­ate stat­ist­ics or math­em­at­ical eco­nom­ics, but there is a col­lec­tion of great aids in the appen­di­cies to the volume to help you out and these are well presen­ted. I really could have used these 25 years ago when I was strug­gling through these courses.

All in all, this is a very good book. It actu­ally does more than it prom­ises and deliv­ers a com­pre­hens­ive and effect­ive course in data ana­lysis with superb hands-on exer­cises to drive home the learning.