Find solution
What are some good "beginner level" data modeling/analytics approaches to kick start a data science/analytics team?

What are some good "beginner level" data modeling/analytics approaches to kick start a data science/analytics team?

  • So I have recently joined a brand new data analytics team at an Ecommerce startup, and I have to decide how to start showing our worth to the rest of the company. What should we start with? Churn Analytics? Propensity models? Micro segmentation for our email receivers? What can I do which is possibly a low-hanging fruit which will also "wow" our stakeholders? Any suggestion is helpful. Thanks!

  • Answer:

    ATA It sounds like you have medium-sized data, good understanding of it, and pretty good data engineering and stats knowledge. That's great, so you can skip past all of the getting your data online and learning the tools. I'd encourage you to next identify an investigative analytics use case. That is, rather than productionize something, focus on refining a model of something that would result in an insight or change you can describe to people. Yes, churn prediction might be a good quick win. It is not necessarily a big-data problem since the training examples are relatively few and coarse. You can probably look at per-user, per-day behavior and do well, and there's not that many. That's good because you can throw R or scikit at the problem. Next I would identify an offline operational analytics problems to solve. Put the static churn model you computed into production, query it at runtime to intervene with an offer or help for "high risk" users. Measure the effect.  Consider implementing a multivariate testing framework or product so you can run experiments and check the effect on behavior after the fact. Last I'd look to implement a real-time operational analytics problem, like a recommender system. That's the canonical example. That's more complex from an engineering perspective for sure.

Sean Owen at Quora Visit the source

Was this solution helpful to you?

Other answers

Churn modeling is a good one. I'd also recommend some sort of a segmentation model that breaks user groups down by acquisition source. Both of these things can be visualized pretty easily, which helps with "wow factor". I wouldn't bother with email re-engagement analysis -- the incremental improvement numbers are so small (<1-2%) that it's difficult to really impress upon people how big of an impact the results can have.

Eric Benjamin Seufert

If you are at an E-Commerce startup, here are the business problems you could address using analytics: 1) Recommendations (on product pages, checkout pages, landing pages, and emails) using a collaborative filtering algorithm 2) Clustering models to segment your customer base holistically (based on transactional behavior, attributes, email engagement, discount-seeking behavior, etc) 3) Response models to predict response to email campaigns of different types 4) Retention/churn modeling to predict which customers are likely to churn 5) Model to predict lifetime value These are in the descending order of perceived value (in my humble opinion).

Raj Bhatt

For a startup, I am assuming you have access to analytics softwares like SAS, SPSS or R, etc. 1) Raj Bhatt has covered up most of the things. 2) Analyze some basic patterns in the data (Exploratory data analysis will give interesting findings and in turn mind blowing insights) 3) You can do social media mining or data mining using Python (most popular) or freely available tools like http://import.io ) to compare yourself with other e-commerce companies. This is complex, but will create 'WOW'. 4)You can also try Google Analytics to track the different trends amongst visitors to your website 5)Advertisement Placements

Amrut Deshmukh

It might be interesting to look not only at churn and lifetime value, but at a more granular level, how are users interacting with your ecommerce site? If an individual purchases something over $50 on their first visit, are they more likely to return? You can also look at marketing efforts, are there certain campaigns or promotions that lead to lower churn or higher average order sizes? Information like that can reveal what your business is doing well and what to replicate.

Kelli Simpson

To quick start a data analysis and data modeling, we have several approaches for beginner level. For the analytic team and data science team, one best tool is use now a day. But it’s not much popular as it should be considering, named http://www.easydatafeed.com If you are not a professional in this field, if you are at beginner level then need not to worry. You can easily analysis on the data modeling approaches.  Check for hackers, data scraping speed, conversion of files, split files, zip files are features of Easy Data Feed. So you can easily and quickly analyze on data mapping and data modeling approaches. If you want to know more about this tool .You can go to here: http://www.easydatafeed.com/open-source/ They also have developers you can hire to do the job for you, their Skype is “easydatafeed”

Brad King

I just found this course recently by Jigsaw Academy - http://jigsawacademy.com/beginners-course-analytics/. It is a beginners course for anyone who has no prior knowledge to analytics. The prerequisites for the course is knowledge of Excel, with logical and reasoning skills. Must try

Anonymous

Related Q & A:

Just Added Q & A:

Find solution

For every problem there is a solution! Proved by Solucija.

  • Got an issue and looking for advice?

  • Ask Solucija to search every corner of the Web for help.

  • Get workable solutions and helpful tips in a moment.

Just ask Solucija about an issue you face and immediately get a list of ready solutions, answers and tips from other Internet users. We always provide the most suitable and complete answer to your question at the top, along with a few good alternatives below.