,

Thoughts on Distributed Data Pipelines – Spark vs Kubernetes

Data Pipelines – Spark vs Kubernetes, or both?

Data gets bigger and teams want to process data faster, what else can you do? There is only so much code tweaking you can do, threads, processes, asyncio, it’s only going to get you so far. At some point you have terabytes of data to process, and it requires a decision about some sort of distributed processing system.

In my experience I’ve mostly used two different distributed data processing systems in production, Spark and Kubernetes. To be honest the choice has always been obvious when to choose one over the other. The data usually dictates which system you choose. I’m sure there are super fans of each system who would argue there’s always a way to do any transform or process on each, but sometimes the point is, which system is setup to easily and quickly move the data from one point to another, and transform it as needed.

Read more
, , ,

You Have to Try This… from io import StringIO, BytesIO

StringIO and BytesIO are perfect for making your Python faster.

Ever heard of something called a File Object in Python? Ever heard of BytesIO or StringIO? Your missing out. It’s easy, fast, and wonderful, in short, it’s the best. For some reason IO streams are a totally underused feature that rarely comes up in most code. We all know that memory if faster than disk IO, this is what I use IO streams for.

Read more
, ,

Concurrency in Python Data Pipelines – Async, Threads, Processes … Oh My.

Concurrency in Python is like the awkward group of middleschool boys gathered in the corner of the dance floor all pointing to the girls in the other corner. Everyone talks a big game, pretending like they totally understand the other group and could easily handle the pickup if they wanted to. Mmmhmmm.

When push comes to shove and you actually have to pick the girl to take to the dance floor, it all the sudden becomes this wierd and strange shuffle of interactions that sometimes works out, and sometimes doesn’t. That is Python concurrency in data pipelines for most folk.

Read more
, , ,

The 10 Commandments of Data Pipelines

When building data pipelines all day long, every day, every year, ad infinitum, suprisingly I have managed learn some things. You see the same problems with data pipelines many times over. Years ago it was SSIS (I’m sorry you still have to use it, it just isn’t cool enough anymore), now if it’s not Streaming it must be wrong (Insert eye roll). The technology and what’s hot is always changing, but the 10 Commandments of Data Pipelines never change.

What are the 10 Commandments of Data Pipelines that thou shalt not break? Glad you asked.

Read more
,

Data Engineering – A Day in the Life.

Yeah, data engineering seems to be a hot topic today, as much as data science is/was 3 years ago. What does a data engineer do, what skills do you need? Peruse the job postings it will quickly become overwhelming. Spark, Hadoop, SQL, Python, Scala, ETL, Data Warehousing, various Data Sciencey Things, Streaming, Analytics, Business Intelligence, Machine Learning, blah blah blah. What are the top skills need to be a successful Data Engineer? What does the average day look like for a Data Engineer? Here is my two cents, it’s probably worth what you paid to read this.

Read more
, ,

Web Scraping + Sentiment + Spark Streaming + Postgres = Dooms Day Clock! PART 1.

So after watching way too many end of the world movies on Netflix I decided the best way to prepare for the Zombie Apocalypse would be to give myself a way to know when the dead are about to crash through my living room window (while I’m eating popcorn watching zombies on on Netflix of course). This is one reason I love Python, I knew I would barely have to write any code to do this. I figured if I could scrape the popular news sites and do some simple sentiment analysis, get the government threat levels, some weather alerts etc, jam all this data together I would get a perfect Dooms Day clock to tell me how close we are to the end of the world on any given day. So lets begin. All the code is on GitHub. Here is visual of what I wanted.


Read more
, ,

Please Sir, May I Have Some More Parquet?

I’ve been wanting to follow up on a post I did recently that was a quick intro to Apache Parquet, specifically when, where , and why to use it, maybe test some of its features, and what makes it a great alternative for flatfiles and csv files.

Read more
, ,

Columnstore Indexes – Always Faster Uh?

Columnstore indexes promise to be the savior of every data warehouse. So, what are they, when should you use them, when to stay away? Columnstore indexes are just what they sound like, data physically stored in a columnar way. This is what makes them so fast when it comes aggregating large amounts of data. The data is compressed and similar values are stored together, the database engine can grab all the values it needs to SUM for example, very quickly, this all leads to faster query results.

Read more
, ,

Hadoop and Python. Peas In a Pod?

Last time I shared my experience getting a mini Hadoop cluster setup and running. Lots of configuration and attention to detail. The next step in my grand plan is to figure out how I could use Python to interact ( store and retrieve files and metadata ) with HDFS. I assumed since there are beautiful packages to install for all sorts of things, pip installing some HDFS thingy would be easy and away I would sail into the sunset. Yeah…not.

Read more
, ,

Look at my Hadoop Cluster… oh, and Don’t Sneeze.



I’ve been wanting to get more hands on experience with Apache Hadoop for a years. It’s one thing to read about something and say yeah… I get it, but trying to implement it yourself from scratch just requires a whole different level of understanding. There seems to be something about trying to solve a problem that helps a person understand the technology a little better.

Read more