What advice would you give to a newbie in analytics?

#1

I’ll start!

#My2Cents - Before you do any kind of analysis, SPEND SUFFICIENT TIME CLEANING AND STRUCTURING YOUR DATA.

I’m sure everyone here will agree that data accuracy is imperative for action-worthy analysis. Data cleaning is often labelled as the “unsexy” part of the job but everything is waste until the accuracy of data is ensured. User behaviours are changing so rapidly these days that the timeliness of the data is also equally important.

As a freshly minted analyst armed with a folder full of certificates, one could be aching to use all kinds of fancy charts, scripts and ML algos. However, all that jazz will only make good music when the instruments are perfectly tuned.

3 Likes
#2

I’ll follow!

Before you even start developing your dashboards, reports and other analysis, you should understand the business and its requirements. So, surround yourself with somebody who knows the business quite well in order to understand it. This approach will save you sometime, sparing yourself from messing around with the database schemas to perceive where the data is!

2 Likes
#3

Couldn’t agree more, Luis! Understanding the business and its requirements brings more context to the process.

#4

Agree with you both, data literacy in the context of your business operations is absolutely crucial.

Also I will add - spend time understanding the core/source systems you are extracting data from. What is the connection between inserts/updates in the core system and the business processess.

Also I will tell any newbie, please be consistent with naming standards (of columns/measures/hiearchies, parameters, visualizations etc) and provide good documenation :slight_smile:

5 Likes
#5

Good documentation!! So important!

#6

Don’t be afraid to ask questions and get feedback from others. This applies to both “how do I…?” questions as well as “do you understand this information if I present it this way?” Some of our best dashboards have come from a sort of mob-programming, where everyone has had input into design ideas as well as making sure the data is actionable and not just nice to have.

4 Likes
#7

Asking questions to make sure data is actionable. Love that!

1 Like
#8

If you have one, run your ideas by your team. They are there to help and most likely look at a problem differently and may have an addition to your solution.

3 Likes
#9

Team work makes dream work. :raised_hands:

#10

Just because you can, doesn’t mean you should!

Keep your reports clean and only show the necessary and relevant information.

4 Likes
#11

That’s so well said, David! It can be tempting to use fancy visualizations to make the reports look flashy but if it doesn’t convey the necessary information easily, the purpose is defeated.

1 Like
#12

I posted this article for new users, and I feel so strongly about it that I’m risking duplication by posting here. I see this ‘recreate our reports’ mentality all the time, and it’s a surefire way to create something people don’t need.

1 Like
#13

It’s a good read, Jeff, and you do make a compelling case on treating dashboards more than just “pretty reports”. You’re so right when you say that it’s about adding value to the data consumer rather than dumping data on them.

#14

totally agree, Short and sweet!

1 Like
#15

Give your data measures/dimensions meaningful names from the get-go. Trying to go back and figure out what Field1 is wastes time and adds to frustration.

2 Likes
#16

Oh, that’s such an important tip! Dimension names should be self-explanatory. Supplement that with a well-defined data dictionary and we have a winner!