The Land of Oz Ozzie Liu

Thoughts from the Final Week of Metis

So this is the final week of Metis and I’m busy working on the final project with crossword puzzles, but I also had some time to reflect of my journey so far.

I’ve put in a lot over these 12 weeks, worked really hard, learned a lot, and have some great projects to demonstrate. But at the same time I feel like I know less about data science today then when I did coming in. Sounds bad but it’s good. Here’s what I mean.

The More I Learn, the Less I Know

I mean that Metis has broaded my horizon of the data science field. There’s an expansion in the breadth of what I need to know. You may have seen these venn diagram and subway maps of data science skills.

venn The Venn Diagram that show data science as a intersection of some important areas

metro Metro map of the skills and tools that a data science will need

But what I’ve found is that the venn diagram needs to have more areas, like visulization, story telling, and business perspective. The subway map needs to have more lines and stations. A data scientist needs to be familiar with a lot more than just machine learning or data mining. There’s data engineering, visualization, front-end, javascript, database with SQL and NoSQL, business application, data structure, algorithms, natural language processing, just to name a few.

So just off the top of my head, I’ve identified these areas:

  • Programming
    • Python
    • Data structure
    • Algorithms
    • SQL, NoSQL
    • Hadoop: Hive, Pig, Spark, Scala
  • Machine Learning
    • Supervised
    • Unsupervised
    • Streaming
  • Data Engineering
    • Ingestion
    • Wrangling
    • ETL
    • Storage
    • Security
  • Statistics
    • Designing experiments
    • Measuring effectiveness
  • Visualization and story telling
    • Presentation and communication
    • Web app (Javascript, MEAN, D3)
  • Business perspective

There’s so much more I need to master!

Repercussions

But to me, this means a few things:

  • I’m humbled: When I started Metis, I thought knowing some machine learning, big data, some math and statistics and programming would be enough. And I was humbled. I had a lot more to learn.

  • I’m challenged: None of these areas are trivial. But I love learning and I love being challenged. I really enjoy figuring out and implementing new things. So even though 12 weeks is not nearly enough to understand everything, I will soon have the opportunity to digest everything at my own pace.

  • I’m excited: This is very exciting! I’m really passionate about data science. And I want to be able to pick up these new skills to solve problems.

  • I can be part of a growing field: I still believe that data science will continue to grow. And I’m joining it at a very exciting time! There’s a lot of opportunities to use data for business, but also to contribute to open source tools!

Where do I stand now?

After 12 weeks at Metis, I know I’m much stronger in programming. I understand and can utilize data structure and algorithms effectively, especially when it comes to large scalable data sets. I also learn how to deploy a full stack project in data science from database to front-end web app. And I learned the chops in visualizzation and presentation.

In the next few days, I’d like to put together a chart to visualize and quantify my skillsets.