…brew your own Business Intelligence

Data Mining and Predictive Analytics Learning Plan

In the last post, I mentioned that I would be putting together a learning plan to facilitate a shift in focus from SSAS into the realm of data mining and predictive analytics. Clearly this is not going to be a quick and easy journey – but I think it is going to be a rewarding one and an important move to make. So, without further adieu, below is my tentative short-term learning plan for building the necessary foundation:

  • Join local R user group
  • Online Classes
    • Udacity
      (~200 hrs or 20 weeks @ 10 hrs/wk)

      • Intro to Statistics (~50 hrs)
      • Intro to Data Science (~50 hrs)
      • Machine Learning 1 – Supervised Learning (~50 hrs)
      • Machine Learning 2 – Unsupervised Learning (~25 hrs)
      • Machine Learning 3 – Reinforcement Learning (~25 hrs)
    • Coursera – Data Science Specialization
      (~110 hrs or 11 weeks @ 10 hrs/wk)

      • The Data Scientist’s Toolbox (~12 hrs)
      • R Programming (~12 hrs)
      • Getting and Cleaning Data (~12 hrs)
      • Exploratory Data Analysis (~12 hrs)
      • Reproducible Research (~12 hrs)
      • Statistical Inference (~12 hrs)
      • Regression Models (~12 hrs)
      • Practical Machine Learning (~12 hrs)
      • Developing Data Products (~12 hrs)
  • Books

I’m sure there will be some minor adjustments here and there but I think it’s a pretty solid start. If you have any suggestions or adjustments, please reach out 🙂

Note: to be clear, I’m certainly not abandoning my deep dive into SSAS and the semantic layer. I love SSAS and think it is a beautiful product that is hear to stay for quite a while. However, one cannot ignore Microsoft’s “cloud first” development strategy which so far has not included SSAS. So I see this as a nice opportunity to branch out into another area until progress resumes.

The other clarification I’d like to make is that I’m not expecting to come out of this endeavor as a professional statistician. The goal is simply to become a valuable resource for data mining and predictive analytics focused projects. To be able to communicate effectively with statisticians and implement solutions in conjunction w/ input from said statisticians and subject matter experts.

Update: Andy brings up a great point in the comments about calculating required hours and duration…so I’ve added hours to the courses based on the estimates from overview page for each class. I’m also expecting a fair amount of sidetracks when I feel a deeper dive is necessary – but this will likely come further along once I start trying to build things 🙂

 

9 thoughts on “Data Mining and Predictive Analytics Learning Plan

  1. PassedBI says:

    not bad ) at your mind, how many hours/week are you going to spend and how long will it take?

    1. Bill says:

      Good question. I’ll probably spend at least 10hrs/wk starting out and see where that takes me. Currently there is not a finish line…so I’ll probably need to re-visit in a few months and reassess my goals.

  2. SQLAndy says:

    Bill, that seems like a sound start to building a plan. I’d like to see the hours the courses list (I imagine you’ll spend more with some offline practice) – is a few months at 10 hours a week realistic? (10 hrs a week is a lot to start with, but nice to know the estimated hour count, even if it does change). There is a nice “done” to finishing these and I like that also, but do you have thoughts on how close this puts you to your final goal? What might the next phase look like?

    1. Bill says:

      Thanks Andy – you’ve made some great points. 10 hrs does sound like a lot but then again, I don’t have any kids and I hate watching sports 😉

      No idea how close this puts me to my final goal. This is just the closest peak at the moment. Once I get to the top of this hill, I’m sure I’ll see another mountain with an even higher peak.

      1. SQLAndy says:

        Bill, I always say “show me the time on your calendar”. If you can map out 10 hours a week, that’s good! As far as the goal, I hear you – that is absolutely the hard part, to understand where you on the path. Defining a goal based on course or exam or similar is sound, but your real goal – can you say it something like “I want to understand the core concepts around x” or “I want to be able to do X” or “I want to be able to do an interview for an entry level position doing x”? That helps you think about if the plan you have is enough. I’d also suggest seeking out someone who already does this kind of work and show them the plan, get some feedback.

  3. Craig Kilgo says:

    Is there anything on the MIT OpenCourseware that would fit in well?

    1. Bill says:

      definitely one source I didn’t consider. I’ll check it out – Thanks Craig!

Leave a Reply