R-Services, Quick Access to R Data Mining from SQL Server Database

I just finished the Data Science End-to-End Walkthrough tutorial. I’m finally getting around to investigating R-Services. From what I’ve seen so far it’s very easy to get started with.

Using SQL Server R-Services I can leverage:

• in-memory querying
• Columnstore indexing
• RevoScaleR which is good for large datasets and parallel algorithms
• “Extensibility architecture”

This is awesome because I can essentially use SQL Server’s computational power for running my R code.

Notes about Microsoft’s Tutorial

The introduced me to the combination of using R code and SQL server functions like stored procedures with data on my SQL Server to perform data analysis more efficiently than if I was importing the raw data into R itself.

In the tutorial we used the R code, SQL data, SQL function combo to develop and test a classification model for some NY taxi fare and tip data.

Takeaways

• Using SQL server data I built, trained and tested a model
• Then I deployed the model to SQL Server so that the script runs faster using all of SQL Server’s “magic”
• I could reduce time in completing analysis by working on analysis models with a subset of data
  • Then when the model is acceptable I could deploy it onto SQL server as a stored procedure
  • I could repeat this process over and over

What’s Next

• I won’t be working on very large sets of data for work for a while
• I’ll instead continue to focus on transferring all my existing data to SQL Server and performing data analyses really fast on my small datasets
• I also want to make lots of stored procedures utilizing R code so I can plot very quickly
• Then as I scale up in terms of data size or complexity of models I’ll have the powerful SQL Server R-Services background

I probably won’t be able to get back to R-Services stuff too much in the next week because I am trying to focus on scripting in Python because I need to free up time.

Overview

Injury shapes playoff runs in the NBA. I want to use data analysis to see if I can find some interesting trends related to player injury. I wonder can NBA injury analysis relate to utility asset management? Can I group players and teams in a way similar to make and model type for assets?

Honestly, I’m just really bored with my current data analysis education. I think it’s because data is boring if it’s not something you enjoy or are a fan of. I’m definitely a fan of the NBA. I enjoy playing basketball. I’m hoping that relating to each line of each table and each data set will allow me to dig deep into this analysis and learn. Even financial data sucks. Maybe because I’m not working hard enough to make enough money lol. But the NBA is forever.

What I Want to Do with NBA Injury Analysis?

I want to figure out how injuries affect team accolades as well as individual accolades. How many injuries do championship teams have on average? If a team’s injury count passes that threshold how does their probability of winning a championship change? How can you rank the effect a injury to a certain caliber player has on his team’s success for the season? For the decade? For the history of the franchise?

How do injuries affect the trajectory of a player’s career?

I think this could be interesting for my job because we are interested in asset management statistics.

Thanks to protransactions.com for the data.

Interesting Stuff So Far

Top 10 injuries (causing DNPs) of all time!

Marcus Camby had a hard career.

Obviously I’ll have the do some error checking before I really get into this but I’m excited to get started. We’ll see how I really do. I stop doing a lot of things abruptly smh.

Plot to Staht

I’m slowing it down today.

#toomuchsauce

#toomuchlow-fatlabeledfood

#toomuchkeylimepie

Top Binges (viewer discretion advised)

For April this year I’ve eaten 2789g of sugar. That’s about 90g of sugar a day for the month. And I haven’t even eaten today yet. Sheesh.

My highest sugar intake day this year was 239g on March 31st where I ate half of a key lime pie. (The day I ate the other half, March 30th, was a 139g day)

The second date this year was 228g on April 14th. Macaron food truck..

Third date was 227g on April 28th. Another half key lime pie. (196g day finishing the pie on the 29th.)

I have a problem with key lime pie. I don’t know where it started.

What would be the “best” macro to recover from a binge of? (Here is an interesting article about cheat-mealing “smarter”)

Fat-Free Not Too Heal-thy?

Besides my binges, another thing I thought I was doing good eating with was eating fat-free Greek yogurt. I thought the low carb, low-fat macro combo was making it a super-find. But what I’ve learned is that for most fat-free foods they replace the fat with sweetener to give the consumer a more satiated feeling. I am going to the supermarket now to compare the low-fat and full-fat versions of my favorite brands to see the sugar difference.

What’s the Plan?

Here’s to learning. I’ll miss you sugar (and probably come back to you a lot in times of weakness)

I’m going to make an effort of having smarter binges/cheat days/days of weakness by trying to limit the sugar and prepare my body.

I feel like I move from one addiction to another. What will my next one be? I want it to be “producing high quality code”.

That doesn’t sound like it tastes good.

Overview

Next week is an important week for my job team,

We’re visiting an important member to help facilitate discussion, finalize plans for different projects and to get raw data.

Beyond all the technical tasks of the week, my focus is to practice being a good team player in terms of helping out with the non-technical tasks of the week. I often get caught up in my own day to day work activities and feel like I’m being distracted from those tasks in situations like next week. All types of impromptu tasks will pop up like:

• Printing things out
• Directing traffic
• Paying bills
• Taking notes
• Late dinners

etc. but I’m telling myself today to embrace the impromptu situations.

Why I Want to Improve

I want to be a better all-around professional. Even though I cringe at certain professional situations now, because they are new to me, I’d like to soon be able to perform all of my technical work duties while simultaneously being able to handle non-technical, but highly important, professional tasks like:

• Small talk
• Working a room
• Impromptu speaking about projects
• Introducing colleagues to other colleagues
• Introducing myself to new colleagues
• Having an active and useful presence in strategy meetings

I know that’s a lot. A part of me still feels negative towards these skills even though I know I need to improve them.

I ask myself:

• “Why?”
• “Shouldn’t I just get better at technical skills?”
• “Are speaking skills and hospitality skills so important for me?”
• “How do I come across as genuine when I might not want to entertain at the time?”

Right now I think the answer is because as a salesman, a.k.a being able to have someone hire me to do work for them, I need to be able to set the stage for myself to entice the customer and give them confidence to feel confident about me working for them.

It could be more complex than that but right now this answer works for me.

Time to Be About It

So next week my goal is to really focus on the non-technical skills. I will make an effort to not think of these tasks as troublesome or a waste. I will think of them as real on the job training preparing me for future growth.

and trying to post regularly again.

mission accomplished.

oh and kobe made a poem about Family Matters

GOAT!