Category: Tech

Analytics From Day 1: How to Ensure Smooth Data Integrations with New Production Systems

Sam Bail
Birds-eye view shot of roads

This is a pre-print of a post that will be published on the collectors.com tech blog.

We recently went through one of our most epic product launches at Collectors: Customers are now able to submit cards for grading to PSA, add them to their collection on the new collectors.com site, and choose to have their valuable cards stored in an actual physical vault. Oh, and all of this will be accessible via a single login, their “Collectors ID”, which replaces the multiple logins customers previously had to manage across our different products and business units like PSA and PCGS. While all these new features are integrated to provide a seamless experience to our customers, we’re dealing with a significantly more complex architecture of multiple new systems, databases, and APIs on the backend. Naturally, as with any new product launch, our product managers were keen to get analytics about the use of these features from day 1: How many users actually converted to the new “Collectors ID”? How many items have customers submitted to the Vault? Who is using the new collectors.com “My Collection” feature? 

In order to provide these kinds of data insights right from the go-live, we had to coordinate with several engineering teams to get our hands on the right data and integrate it into our data warehouse. In total, we ended up pulling data from systems sitting on top of four different production databases that were being launched at the same time. Considering how many different systems and databases we were working across, the integrations went pretty smoothly! Within a day of the go-live, I had produced a few dashboards with key metrics that the product and business stakeholders started using immediately to track uptake of the new services. The road to getting there wasn’t straightforward though and involved some amount of scrambling, knocking on different doors, and a few small surprises during the go-live. In this post, I will share some of my lessons learned from integrating with a new production system when you’re looking to provide analytics from the get-go.

Logistics

I’m big on keeping running docs with notes from my conversations and findings when working on a project – I always say I outsource my brain into a Google doc. Keep a doc with (datestamped) notes and “to do” items for every piece of information you find, open questions, as well as a list of who’s responsible for what on the product, e.g. product managers, engineering leads, project managers, etc.

In addition to the running notes, connect with the business stakeholders (product managers, analysts…) early on to document a set of desired metrics along with their priorities and timelines: What do we need to know from day 1? What can wait until some time after the launch? This will also be helpful when exploring the new data models to determine what is actually being captured and what data points may not be available to calculate the required metrics.

If there are standing meetings for the engineering team that’s responsible for the database setup, I strongly recommend regularly sitting in on those meetings. Even if you don’t always understand everything that’s going on, it’s helpful to have the context of what the team is focusing on, and establish a relationship with them. As data engineers, we’re often pretty removed from our counterparts on the data producer side, but knowing the people on the team (and having them know you) can be helpful in working together more effectively. 

Infrastructure

Once you know who your engineering point of contact is, the first question you’ll want to ask is: How do we get access to the data? Assuming we’re talking about data that lives in a relational database, here’s a short check list of information you need to get from the engineering team that’s responsible for the database setup:

  • Find out (and document) what cloud service the database is hosted in
  • Will you get access to a production database or a read-replica? And what permissions will you get, read-only, or will you be able to create temp tables or views if they’re needed by any of the tools in your pipeline?
  • Will there be dev and prod environments? What’s the timing for these being available?
  • How do users and services authenticate against the database? Do we need personal and/or service accounts to log in?
  • How will the logins will be shared? Will you need access to a shared password storage?
  • Do you need an SSH tunnel setup to access the database from any of the tools in your data stack?

It’s best to try and get all these details ironed out as early as possible, since especially tasks like setting up SSH tunnels can take some time. Make sure you can access the database as early as possible to avoid surprises later on, even if there is no meaningful data in there yet. 

Data model

Now that we’ve covered physical access to the data, let’s take a look at things to consider when you’re working with a new data model. I got looped into the production database design process early on and was able to provide input on the data modeling (see also: establishing a good connection with the upstream engineering teams! They’re your friends!). This ensured that the data would be suitable for our data extraction tool (Stitch) and contained all relevant data. Again, assuming you’re working with a relational database, here are some questions you’ll want to cover when talking about the data model:

  • Where is the data model documentation and how is it being kept up to date?
  • For any fields containing value sets, such as status codes, where are the corresponding descriptions stored? Will there be lookup tables in the database, or will these only be stored in code? The latter means you will need to be able to access the up-to-date list of lookups through your infrastructure, e.g. by querying an API (or simply reading the API documentation).
  • Will there be JSON columns? What is the schema for those?
  • What are the constraints on each table and column, e.g. foreign key relationships, NULL values, default values?
  • For datetime fields, will they be stored with timezone (they should)?

Application and data flow

Perhaps most importantly, when trying to make sense of data coming from a production base, we need to understand what the flow of the application is: What workflows (user-created or automated) in the application modify the data in what way? This is absolutely crucial to handling the data correctly and drawing the right conclusions from it. For example:

  • How and when is a record created, and what fields are populated through what input?
  • What workflows cause records to be modified in what way? And what metadata is there to track modifications, e.g. a “last updated” timestamp?
  • Will update timestamps for events such as status changes be tracked in separate fields? Or will there be kind of changelog table that captures these kinds of changes? This also trickles down into your data warehouse models, where you might need to start tracking status change dates right from the get-go.
  • How are deletions being handled? Will there be “hard deletes”, i.e. the record is simply removed, or “soft deletes”, i.e. the record has a “is deleted” or “deleted timestamp” field. And, along the same lines, is there a data retention policy that means data will be dropped or archived after a certain amount of time?
  • If the application is replacing a legacy application, will data be migrated? How do you recognize migrated data? Will there be any gaps or differences between migrated and newly create data?
  • Will there be realistic dummy data (i.e. data that adheres to the constraints and workflows described above) to develop our data models and metrics against?
  • Is there any chance of any test or dummy data getting into the production system? And if yes, how can we recognize and filter for it?

Ideally, your engineering and database admin teams will already have a “best practice” guide for designing new databases, which usually answers a lot of these questions. Otherwise, this might be a good time to start collecting these kinds of design decisions into a guide and encoding them in setup scripts where possible.

And finally…

I hope that this post has provided you with a starting point for a checklist for your next production data integration. All the questions I’ve covered in the above paragraphs should be treated as conversation prompts to elicit existing design decisions, or help influence decisions that are yet to be made. There will likely be some oversights (I have yet to work with *the* perfect production database), but coming prepared with a plan may help you catch some of the biggest issues to getting a good data integration early on. And even with the best preparation, you can probably expect to make some tweaks after the application go-live to adjust to some last-minute database changes or correct some assumptions you’ve made about the data. Developing against an empty data model or even dummy data can be challenging, and you might not nail everything at first try.

One last thing to keep in mind: As data consumers, our downstream use case will most likely be of lower priority than getting the production system stood up – and that’s totally okay. While I would love for data to always be a first-class citizen, I believe it’s pretty obvious that producing a stable production system needs to take priority, and we just need to accept that resource constrained engineering teams may move slower on supporting a data integration. This is why you’ll want to get started early and get these kinds of tasks and questions on the engineering team’s radar as soon as possible.

CC-licensed photo by Ian Beckley: https://www.pexels.com/photo/top-view-photography-of-roads-2440013/

Building a data platform from scratch at Collectors: A tale in three parts

Sam Bail

I wrote an epic blog post series about my experience building a data platform from scratch in my new job, using the “Modern Data Stack” (well, at least parts of it). The post is an account of my first six months at Collectors building a data platform. It is part memoir, part instructional manual for data teams embarking on a “build a data platform” journey. I figured this might be relevant for some of y’all data engineering folks and/or “data teams of one”, so check it out here: Building a data platform from scratch at Collectors: Part 1 (parts 2 and 3 are linked from the post).

Image credit: “under construction” by Pedro Moura Pinheiro is marked with CC BY-NC-SA 2.0.

Don’t be that person. Or: How to not be a Kool Aid Man in the “extended workplace”.

Sam Bail

Have you ever been out to a restaurant or bar with someone you considered a friend, or maybe a partner, or a date, and it turned out they acted kinda shitty towards the wait staff? Maybe they were unnecessarily impatient, rude, dismissive, entitled, or talking down at people? Or maybe you just witnessed someone acting like that in a public setting and felt some amount of “Fremdscham” (the German word for feeling ashamed for something someone else is doing) ? Yeah? That’s because acting like that is generally considered “bad behavior” and most folks are aware of the rules of common courtesy when interacting with other people, usually those in a position of delivering a form of service.

Cool, Sam, but why are you telling me that? Isn’t this like, a tech blog of sorts?

Well, I recently participated in a number of virtual tech events where I witnessed that very same rude, dismissive, impatient, disrespectful, and entitled behavior (yes, this post is a bit of a rant!) from participants towards the organizers and presenters, and it appears to be more of a systemic problem than just a few individuals being annoying.

Here’s an example from a free live training session I recently attended that was the catalyst for this blog post (note the timestamps for the correct order):

The presenter had clearly explained and demonstrated two free options for using the software at the beginning of the hands-on part, and the teaching assistants in the course had responded to every single one of the participant’s questions. And yet, he posted himself into a rage and acted like a complete ass. I can’t imagine that he’d act like that around his office – and if he did, I hope the company would tell him very clearly that’s not acceptable behavior.

(As an aside, another participant joined the live training 20 minutes before the end of the 2 hour session and demanded someone explain to them how to get started. The training was definitely interesting.)

Another example for interactions that are not necessarily disruptive but just look bad are folks asking for help in Slack channels. I just posted about this on Twitter a while ago:

Screenshot of a tweet saying: "I swear every single #general Slack channel in tech is like - User A joined the channel - User A: "HEY GUYS here's a huge stack trace help me fix it for free and asap" Could the eng managers of this world PLEASE sit their engineers down and teach them some manners?"

I’m in a quite a few tech Slack channels and I used to be a maintainer of an open source project, and the typical behavior I notice is:

  • New user joins the channel
  • Immediately posts a question asking for support, often dumping an entire error stack trace into the channel with no warning
  • Frequently cross-posts the same question in other channels
  • Occasionally posts several “anyone?” type follow-ups
  • (Rarely) posts some annoyed or frustrated comment when they don’t receive help
By Source, Fair use, https://en.wikipedia.org/w/index.php?curid=35275158
“HELP ME”

Maybe I should care less about these kinds of things, but man, seeing this is annoying. I’ve muted most Slack channels I’m in because of too many Fremdscham-inducing interactions. Especially in open source communities, this sort of Kool Aid Man behavior (kicking down the virtual door but going “HELP ME” instead of “OH YEAH”, you get the idea) makes you wonder where people left their manners.

Another version of this is the “mouse asking for milk” behavior, which often follows Kool Aid Man behavior once someone receives help. For those that don’t know, the popular children’s book tells the story of a mouse that receives a cookie, then proceeds to ask for milk (to go with the cookie), a straw (to drink the milk), and other favors. This often has the effect of pressuring the helper to dedicate more time and implicitly puts the responsibility of solving the issue on them instead of the original question asker: “If you don’t continue to help me, you’re letting me down and I can’t solve this problem”.

Look, I understand that we’re all trying to get to results as quickly as possible. Fixing bugs and production fires, figuring out a configuration after banging our heads against the wall for hours, trying to get something to work while following along with a live instructor, all these things are annoying and stressful and make us impatient and want HELP. NOW. But we always have to keep in mind that the people on the receiving end are also just… people. Who are usually trying their best to be helpful, but they might have their own stressors, deadlines, time schedules to stick with, and might not have the capacity to drop everything and help. And maybe you’re the one who’s causing the thing to not work (if you’re in tech you’re guaranteed to have had that experience) – might be time to take a step back and take a break.

I’d also like to clarify that I’m not talking about obviously “bad” or illegal behavior. While many meetup groups, conferences, and open source projects have a Code of Conduct, most of the behavior I refer to is not necessarily a violation of a Code of Conduct, but just generally unpleasant. But keep in mind, just because it doesn’t go against any of the rules doesn’t mean it’s not disruptive, disrespectful, or just plain annoying to the organizers, presenters, volunteers, and other participants. And it makes you, and potentially the company you represent, look kinda bad.

How to not be “that person”

So here’s a thought for folks attending any kind of (virtual) events or participating in Slack communities, message boards, Reddit, GitHub conversations, and other communication channels. I don’t know if anyone’s reading this who should be reading this, but here we go. Before posting anything, ask yourself the following questions:

  • Did I read the “welcome” message and instructions of where to post what?
  • Am I posting in the right channel?
  • Is my question clear and can people actually help me based on the information I’m providing?!
  • Did I use the search functionality to try and see if this question was already answered?
  • Am I asking an unpaid volunteer to do extra work? Have I already taken up a lot of their time?
  • Am I being respectful and mindful of people’s time and other responsibilities?
  • Would I post these kinds of things in my company chat, or say it out loud in a team meeting when my peers and managers are around?
  • Can I wait until it’s a good time to ask that question?

And even after posting a question, there are some things you can do to make everyone’s life easier:

  • Check whether someone actually answered the question, or asked for more details. Respond in a timely manner, or at least let them know that you will get back later.
  • Said differently, pay attention and understand that if someone responds to you, they dedicated time to helping you. Be respectful of their efforts.
  • If you don’t get the help you need, well, so be it. Unless you’re talking to the customer service of a service or product you pay for, you are not entitled to receiving any help, like, ever. And even if you’re paying for the service, keep in mind that customer service staff are humans you should treat with respect. Be persistent if you need to. But for goodness’ sake, please be nice.
  • If the problem is resolved, post that you solved it and ideally, share your solution! This will help people later on, and lets people know that you no longer need help.

Tell your coworkers to not be “that person”

And for the managers out there: I know you’re not responsible for how your reports act outside of the work environment, unless that employee is explicitly there to represent your company. But we all know that the workplace implicitly extends beyond the boundaries of your company’s office, Slack, or email, and that employees are often seen as representing the company in the “outside world”, whether that’s good or bad. If your reports or coworkers (or managers…) behave disrespectful or somewhat disruptive (again, without necessarily violating any Code of Conduct) in an “extended work” setting, that’s just going to look bad and quite possibly make people question your company culture and what kind of people you hire. Well, it definitely makes me question what your company culture is like.

This isn’t an easy conversation to have, but I do believe that any company that onboards new employees likely shares (should be sharing?) some form of “rules” of communication, their company values, or other training that usually boils down to “don’t be rude“. It should be easy enough to include that this also applies to external venues such as (virtual) conferences, Slack channels, message boards, meetups, and other spaces in which the employee is present in a somewhat work-related context and may be seen as representing the company.

And for the presenters, maintainers, and volunteers out there…

Hey there, I see you. Well, I am you. I run workshops, teach coding classes, give conference talks, and help out in tech Slack channels. And I know that putting yourself out there and doing stuff out in public, whether that’s as a volunteer or part of your job, always comes with some amount of pressure and anxiety. Dealing with people who are rude or impatient is never pleasant. Here are some thoughts on how to help with this:

1. Set automated welcome messages in Slack and other communication channels explaining to folks where to post and how. Based on my experience, you can expect some proportion of people to actually read them, and some proportion of that to follow the rules. There will always be people who don’t pay attention, but you can make sure that the rules are actually enforced through gentle reminders: Ask your staff or volunteers to nudge people to post in the right channels, which (hopefully) also will be noticed by other members who will help with that. The dbt folks are pretty good at directing their Slack traffic to the right channels using welcome messages and periodical friendly reminders, see the screenshot below.

Screenshot of the dbt Slack channel stating some rules for what to post where.

2. Add an “FAQ” page to your organization’s website. Reshama Shaikh, a data scientist who’s incredibly active in the NYC tech community I’ve been lucky to collaborate with for years now, recently pointed me to the FAQ page of Data Umbrella, a volunteer-led community group she founded. The FAQ cover a range of questions such as “can you give me career advice” and “can you help me find a job” and kindly point out that the group is entirely run by volunteers who give up their free time and pay out of their own pocket for any kind of expenses (such as MeetUp fees).

3. Have a slide on “How we communicate” rules at the beginning of a talk or workshop. In addition to highlighting the Code of Conduct, you can remind people when and how to ask questions, to use the search function, mention that the talk will be recorded and how the recording will be shared. If you have helpers or TA’s, ask them to enforce those rules, e.g. by posting reminders to hold questions, that the talk will be recorded, or links to the material.

4. Make technology work for you. Honestly, this might be a little dramatic, but see item #1 – there’s going to be a certain number of people who don’t read the rules. One way to make technology work for you, in addition to automated welcome messages, is to lock down the “general” Slack channel to allow only staff announcements, which is a good way to avoid the “new user support question dumping ground” effect. Another option to consider for any kind of live event is to only allow participants of a to join until a few minutes into the event, which avoids people not catching parts and then demanding help 45 minutes into a session.

5. It’s ok to not please everyone. I used to have the “will to please” like a freaking Golden Retriever. But you know what – it’s ok to say no, ignore people, or tell them to wait, for the sake of your own sanity. If someone comes into an event 30 minutes late and you’re a presenter or assistant already juggling several participants, well, maybe the person who came late simply won’t get lucky today and will have to figure things out themselves. Be kind, but firm, and let them know that you won’t be able to catch them up. Sorry. Likewise, if you’re helping someone out in a Slack channel and the mouse asks for more milk, it’s ok to let them know if you don’t have the capacity to help them any further… unless you are working in customer support of course and uh get paid to do exactly this. Otherwise, allow yourself to say no if this is turning from something you enjoy into a chore.

And finally…

I focused a lot on the “don’t make your company look bad” argument in this post, but I think it’s also important to point out that general kindness and respect towards people who dedicate their time to maintaining software, running workshops, giving talks or presentations, should be a given. Whether that’s paid or unpaid, we all need to consistently make an effort to see the person on the other side and man, just give em a break. Chill. Be nice. Accept the fact that sometimes you can’t have it your way. It’s ok. The world won’t end.

Remember when I was British?

Sam Bail
Screenshot of the podcast page with a photo of Sam holding a cup saying "geekgirl"

In case you missed it, I lived in Manchester for 5 years and somehow developed a proper Mancunian accent. Somehow I ended up on Nathan Rae’s podcast “Northology” in 2013, talking about Manchester Girl Geeks, a not-for-profit community group I co-founded a few years prior (they’re still going strong, 10 years later!). If you want to listen to 30 minutes of me being proper Northern, the recording is still online.

Hackathons are more than just t-shirts

Sam Bail

I already showcased a gallery of Hackathon t-shirts I designed for our Flatiron Health hacks a few posts back. Of course, Hackathons are much more than just t-shirts, so earlier this year my fellow engineer Ovadia and I wrote up a 2-part series of blog posts about Hackathons.

Let me know what you think and/or give me some Medium clappy-claps!

What the hell is ‘Crisis Driven Development’?

Sam Bail

After a couple of major production fires on our analytics pipelines that required us to drop everything and push through several migrations, my fellow engineer Zach and I looked at each other and admitted that “this was terrible, but things are actually much better now” – and so, the term “Crisis Driven Development” was born. We shaped up our ideas around the concept enough to talk about it at an engineering all-hands at Flatiron Health, and followed up with a couple of blog posts.

Zach’s post (the first part of the saga) focuses on how to be in a good spot to actually keep pushing forward during production fires instead of rolling everything back, and emerge on the other side in a better state.

The second part that I wrote then takes a step back and thinks through what makes ‘Crisis Driven Development’ so successful, and how we can apply those principles to a regular development cycle instead of a crisis – a controlled burn, so to say. And while none of this is entirely new, I do like how it can introduce a slightly different way of interacting and working together than the go-to mode of agile sprints and tickets. Let me know what you think – feel free to comment here, on Medium, or hit me up on Twitter.

Hackers gonna hack… and wear t-shirts.

Sam Bail

Within my first six months at Flatiron I started helping out, and then later running, our company Hackathons: two days every quarter where the entire team (engineers and non-engineers) is encouraged to get (even more) creative and spin up new product prototypes, try out technologies, liberally fix bugs, or dig deep into our data.

An important part of our Hackathon tradition, along with the Thursday night pizza, are the t-shirts, which I’ve been designing almost every time for the past two years. What started out as a necessity has now become one of my favorite parts about the Hackathon – I get to play around with design tools and see people wearing the shirts I designed almost every day! I’m quite pleased with them, so I thought I’d share them with the world. Here are some of the most recent t-shirts I put together for our Hackathons.

hack7 hack9 hack10 hack11

Getting a Raspberry Pi to run, using Mac OS

Sam Bail

I just installed Raspbian for the second time on my Raspberry Pi (I needed the SD card for my digital camera a while ago…) and had some troubles at first getting the SD card to work on Mac OS because of a “Read-only file system” error. Here’s some instructions, in case you happen to come across the same problem!

  1. Download Raspbian, the operating system to run your Raspberry Pi. I download the latest version here.
  2. Unzip the file you’ve just download – you will get a file with the extension .img.
  3. Make sure the little lock on the SD card is in the “unlocked” position.
  4. Plug your SD card into your card reader.
  5. Check if the SD card is writable: Open the “Disk Utility” on your Mac and click the SD card, then go to the “Erase” tab. If the “Erase…” option (see image) is grayed out, the SD card is not writable. Also, if you try to delete or modify any existing files on the SD card, you may receive a “Read-only file system” error.
  6. I don’t know exactly what causes the “Read only file system” problem, but fortunately I found a solution online:
    Plug in the SD card and go to “System Preferences” > “Sharing” on your Mac. Select “File Sharing” and then the little + sign under “Shared Folders”. Here, you can select the SD card (in my case it’s named “CAMERA”) as a shared folder (see image below). The SD card will now be writable!
  7. Close the System Preferences and return to your Terminal. I got the following instructions from Dag-Inge Aas’ website.
  8. Change into the directory where you have downloaded and unzipped the Raspbian image. In my case I saved the file in “Downloads”:
    cd ~/Downloads
  9. Identify the device name of the SD card by typing:
    df -h
  10. The device name will probably be something like disk1s1 or disk2s1. Then unmount the disk, using the SD card’s device name if it is different from “disk1s1”:
    sudo diskutil unmount /dev/disk1s1
  11. And finally, install the disk image file to the SD card. Make sure the raw disk name, that is the “rdisk1” at the end, is right. It will be rdisk1, rdisk2, etc. depending on the device name above:
    sudo dd bs=1m if=2012-12-16-wheezy-raspbian.img of=/dev/rdisk1
  12. Wait for a few minutes until the image is installed on the SD card. You will then see a message right at the end.
  13. Eject the SD card, plug it into your Raspberry Pi, and off you go!

Stardog ASCII art power

Sam Bail

I’ve just installed the Stardog RDF database for the first time (painless. Download, unzip to some directory, set an environment variable, done. TAKE A NOTE, triple stores.) and on server startup I was greeted with this wonderful peace of ASCII art:

screen-shot-2012-11-24-at-01-52-36

You’ve just won me over.

Edit: Downloaded, installed, and loaded one of the example files with Stardog in 2 minutes. User-friendliness win. Now, if you could perhaps explain what that mysterious “-t D” flag is…

$ ./stardog-admin create -n myDB -t D -u admin -p admin --server 
snarl://localhost:5820/ examples/data/University0_0.owl

Without the flag, the loading simply fails with “Authentication failed”. Eh.

Coding Horror: Please don’t learn to code

Sam Bail

To those who argue programming is an essential skill we should be teaching our children, right up there with reading, writing, and arithmetic: can you explain to me how Michael Bloomberg would be better at his day to day job of leading the largest city in the USA if he woke up one morning as a crack Java coder? It is obvious to me how being a skilled reader, a skilled writer, and at least high school level math are fundamental to performing the job of a politician. Or at any job, for that matter. But understanding variables and functions, pointers and recursion? I can’t see it.

…

An interesting blog post by codinghorror.com’s Jeff Atwood on why not everyone should learn to code. I generally agree with the points he makes (don’t learn to code for the sake of it, and don’t do it for the “fat paychecks”), but I also believe that even just the simplest attempts to learn how to code will give people insights into how computers work. This, in turn, will take away some of the myths surrounding computers (“don’t touch that! It will break!”) and maybe lead to a better understanding of what’s going on inside those boxes – we need not only more good programmers, but also digital literacy of the wider public!

Read the complete post on codinghorror.com