We don’t need more data scientists — just make big data easier to use

60 Comments

Credit: Sergey Nivens/Shutterstock.com

Virtually any article today about big data inevitably turns to the notion that the country is suffering from a crucial shortage of data scientists. A much-talked-about 2011 McKinsey & Co. survey pointed out that many organizations lack both the skilled personnel needed to mine big data for insights and the structures and incentives required to use big data to make informed decisions and act on them.

What seems to be missing from all of these discussions, though, is a dialogue about how to steer around this bottleneck and make big data directly accessible to business leaders. We have done it before in the software industry, and we can do it again.

To accomplish this goal, it’s helpful to understand the data scientist’s role in big data. Currently, big data is a melting pot of distributed data architectures and tools like Hadoop, NoSQL, Hive and R. In this highly technical environment, data scientists serve as the gatekeepers and mediators between these systems and the people who run the business – the domain experts.

While difficult to generalize, there are three main roles served by the data scientist: data architecture, machine learning, and analytics. While these roles are important, the fact is that not every company actually needs a highly specialized data team of the sort you’d find at Google or Facebook. The solution then lies in creating fit-to-purpose products and solutions that abstract away as much of the technical complexity as possible, so that the power of big data can be put into the hands of business users.

By way of example, think back to the web content management revolution at the turn of the century. Websites were all the rage, but the domain experts were continually banging their heads against the wall – we had an IT bottleneck. Every new piece of content had to be scheduled and sometimes hard-coded by the IT elite. So how was it resolved? We generalized and abstracted the basic needs into web content management systems and made them easy for non-techies to use. As long as you didn’t need anything too crazy, the problem was solved easily, and the bottleneck averted.

Let’s dig a little deeper into the three main roles of today’s data scientist, using online commerce as a backdrop.

Data Architecture

The key to reducing complexity is to limit scope. Nearly every ecommerce business is interested in capturing user behavior – engagements, purchases, offline transactions and social data – and almost every one of them has a catalog and customer profiles.

Limiting scope to this basic functionality would allow us to create templates for the standard data inputs, making both data capture and connecting the pipes much simpler. We’d also need to find meaningful ways to package the different data architectures and tools, which currently include Hadoop, Hbase, Hive, Pig, Cassandra and Mahout. These packages should be fit for purpose. It comes down to the 80/20 rule: 80 percent of big data use cases (which is all most ecommerce businesses need), can be achieved with 20 percent of the effort and technology.

Machine Learning

Surely we need data scientists in machine learning, right? Well, if you have very customized needs, perhaps. But most of the standard challenges that require big data, like recommendation engines and personalization systems, can be abstracted out. For example, a large part of the job of a data scientist is crafting “features,” which are meaningful combinations of input data that make machine learning effective. As much as we’d like to think that all data scientists have to do is plug data into the machine and hit “go,” the reality is people need to help the machine by giving it useful ways of looking at the world.

On a per domain basis, however, feature creation could be templatized, too. Every commerce site has a notion of buy flow and user segmentation, for example. What if domain experts could directly encode their ideas and representations of their domains into the system, bypassing the data scientists as middleman and translator?

Analytics

It’s never easy to automatically surface the most valuable insights from data. There are ways to provide domain-specific lenses, however, that allow business experts to experiment – much like a data scientist. This seems to be the easiest problem to solve, as there are a variety of domain-specific analytics products already on the market.

But these products are still more constrained and less accessible to domain experts than they could be. There is definitely room for a friendlier interface. We also need to take into consideration how the machine learns from the results that analytics deliver. This is the critical feedback loop, and business experts want to provide modifications into that loop. This is another opportunity to provide a templatized interface.

As we learned in the CMS space, these solutions won’t solve every problem every time. But applying a technology solution to the broader set of data issues will relieve the data scientist bottleneck. Once domain experts are able to work directly with machine learning systems, we may enter a new age of big data where we learn from each other. Maybe then, big data will actually solve more problems than it creates.

Scott Brave is co-founder and CTO of Baynote, an e-tail and e-commerce advisory business. He is also an editor of the “International Journal of Human-Computer Studies” (Amsterdam: Elsevier) and co-author of “Wired for speech: How voice activates and advances the human-computer relationship” (Cambridge, MA: MIT Press).

Photo courtesy of Sergey Nivens/Shutterstock.com

60 Comments

Phiem Nguyen

Thank you, I wish scientist will brake the abstract circle domain on our privacy from mind and physical life. I would like to sale this products.

Jon Robinson

Big data, little data – the key to good analysis, is knowing what question(s) one wants to answer and then identifying what combination of information sources need to be examined to get answers. Along the way, one needs to use appropriate data collection and analysis methods for answering the question(s). Collecting or getting access to large amounts of data without knowing what the question(s) are is not very useful. The most frustrating thing in statistical consulting is when someone comes to you with their data already collected (or in the case of big data, their data streams already procured), asking you to process it to get the answer they want; but on examination, the data is not the correct type, was not collected in the right way etc. to answer their question regardless of how it is processed. Dash boards and other easy to understand displays can be very misleading if there is no understanding of what is behind them. Templates are fine if you understand all of the assumptions behind them and what they do (besides create pretty pictures).

Sudarshan

Amen, somebody needed to write this. Even startups and SMBs have enough data that they’re unable to derive value out of; hiring a data scientist is out of the question – only better tools and technology can help.

Dipesh

Great article. What I particularly like about it is the thought that the “Data Scientist” is no panacea in the face of the scale of the need (the 80% of companies who can’t access/afford a Data Scientist).

In my personal opinion, focusing so much on the Data Scientist could end up recreating much of the same type of analytic (process and people) bottlenecks that we already have today.

nathj

It seems to me that this boils down to a case of reconsidering and refactoring the problem scope. Any problem can be broken down into inputs, processes and outputs. The process generally take you from inputs to outputs. Big Data is being treated differently at present for reasons I can only guess at; it’s new, it’s exciting, it’s mis-understood?
The point being made here is spot on in my opinion, we just need to break the problem down into achievable chunks. Re-evaluate what we are trying to achieve overall and see she big data fits into that in order to get the most out of it.
Ultimately data is data. How we store it, use it, manipulate it is just processes. It is still data no matter how much of it there is or what label we give it.
Thanks for a great article.

marianudo

Many people thought that programmers wouldn’t be needed if someone was able to abstract business generic needs and make a tool that generates code for us. Actually it hasn’t become true. With data scientists is likely to happen the same. We might think that with suitable tools we don’t need them. But big data and scalable architectures are a technical challenge and any customisation (always needed) has to be made by qualified staff. Data Scientists will always be necessary, as well as programmers are even in the most standard industries.

Scott Linford

marianudo nailed it

Let’s write the “one program” that writes all the other programs. It’s a template. It’s a wizard… It’s snake oil. But now repackaged to make sense of unstructured effluence? Right. Are decision makers really this desperate? Enough to buy this nonsense?

laseray

I don’t even have to read past the title to know this idea is flawed. It is the same idea that people had when object-oriented programming was supposed to make everything in software so easy that even end users could just build their own. That never happened, and it couldn’t because it was an instance of silver bullet mentality. And big data today is suffering from the same mentality in some circles. It is a mentality that mistakes a technology for an absolute solution. There have been many instances of this mistaken kind of thinking. The companies that don’t get it eventually meet their demise if they don’t change their point of view. And saying you don’t need more data scientists for data is like saying you don’t need more programmers for software, it just won’t work.

Scott Brave

Interesting counterpoint. I don’t think what I’m proposing is as extreme as that though. Trying to create an abstraction layer for all big data analysis and machine learning clearly isn’t feasible. But, like I responded to Keyser above, the trick is limiting scope. What gives me hope is that there are examples of where it’s worked already: like in the recommendations space.

My analogy is closer to replacing programmers with applications. There have been plenty of needs in the past that required programmers and now only require technology (like ad servers and social networking features). As others in the comments have suggested though, perhaps it’s just a moving target: there will always be new needs for data science. Whether the bottleneck truly loosens up then will depend on how quickly we can keep ahead of the curve with useful product.

Ayman Yacoub

Interesting article, but debatable. Big data requires special talent and breed to work with and analyze to conclude and make final business decisions. It won’t be in the very near future that we’ll see a tool as such that can take care of this ambiguity even though there could be solutions implemented to enhance the process and make it easier to understand.

Jon Bloom

Don’t fix the traffic problem, have fewer cars on the road. Seriously, data has gotten more complex. Would be ideal to simplify, not always easy, but definitely possible. Data storage is easier. Data delivery has gotten easier in form of reports, visualizations, dashboards. ETL is the elephant in the room, very challenging to connect the dots between disparate data sets, and then find insight. And gap between IT and Biz remains, connecting business rules to technology. Either way you slice it, data to information is a HOT topic!

bloodfreak

Here we go again. Business people whining, “it’s too hard! It should just be easy!” But didn’t we just give you your shiny iPad to play with?!? Didn’t we just give you your cloud?!? Stupid nerds had to show off their little hadoop and now the MBAs think distributed number crunching is the panacea for all business problems. But I guess it gets VCs salivating, they throw some money around, people are employed, it’s all good. Yay, Big Data!!!! But off-the-shelf, template-based big data analysis will be old hat in short order, because once it’s in everyone’s hands, it’s table stakes. The only “insights” to be derived will come once again from custom analysis. Your bottleneck never really goes away. The goal post has moved. It’s the march of business and technology. Congratulations, you’ve graduated to a bigger treadmill.

Yong Sheng has probably made the best point here; just because you put the tools in their hands doesn’t mean they will know how to use them, no matter how pretty and easy-to-use the interface is. Business people are notoriously bad at numbers (see “bottom line” mantra).

Bah, maybe the Christmas season is making cranky :-)

S Moffitt

If you look at Big Data as a science, then it appears that the science is undergoing a Kuhnian paradigm shift. If so, there will continue to be a need for data scientists until there is a concensus on what the problems are and how to go about solving those problems. This field is so new that we do not yet know what we can do and how we can use the possibilities that the technology offers. Until we do, there will be a continued need for people to cook bespoke analytical tools or modify products to do what the business wants.

Hanjra

one can create products to drill down the data, but is the drilling really giving you the right direction to optimum result? An anlytical product can simplify the data for you but a data scientist will infer and produce what can go away unattended..

Steve Na

.
In my opinion, data scientists will be in demand for a long time.

1) who will monitor the systems that make big data easy to use for the domain experts? How do you know that the models that are being generated are correct and that the product is working properly? At minimum, a data scientist will be needed to develop such products, but also to deploy and/or monitor these types of products. Who will check to see if the domain expert is using the system properly and interpretting the results correctly? Who will be there to answer questions when the domain expert has them?

2) Products are fine for fast followers, but for the market leaders, 80% is not good enough. To use online commerce as an example as the author does… Amazon will not use Baynote for product recommendations. They need their own data scientists to build custom solutions that get 100% of the job done. For other companies that compete with Amazon that are smaller and trying to narrow the tech / science gap with Amazon, using something like baynote makes complete sense and I agree there is less need for data scientists in these smaller companies. But, if you are amazon, I don’t see any time where they will not employ data scientists.

3) Yesterday’s innovations become products (product recommendations, personalization, etc). But, the time between when data-driven innovations happen (by humans) to when they become products will provide the opportunity and demand for human data scientists. And so, unless we think ecommerce is done innovating, then I think human data scientists will continue being in demand even in companies that prefer to purchase productized solutions since they don’t want to be left behind.

4) Online commerce was used as an example space where data scientists can be replaced with better products that allow domain experts to use big data without data scientists. I think this breaks down when you look at other areas. Facebook, LinkedIn, Twitter… Google, Yahoo…. Zynga, Playdom…. etc, etc. All these companies hire for data scientists. If they were to replace their data scientists with a product, I don’t see what product that would be or what vendor would fill that gap. In search, there are 3 big engines, what vendor will make a system that replaces what Yahoo and Google have built themselves? For all these companies, data science is fundamental to their success and differentiation so I don’t see how they can use some product instead. If the products are going to be built internally by these companies and not purchased from a vendor, then these companies still need to hire data scientists to build it… probably a lot more if they want the system to be run by a domain expert with very little interaction with data scientists. That seems like a lot of work just so a domain expert doesn’t have to talk to a data scientist. It would be a lot easier to build a system for data scientists and have them work together with domain experts… which is what is happens right now.

Keyser

Disagree! Depending on the domain, the data, and the application the feature extraction process as well as the task at hand require expert analysis & cannot be done automatically. If you don’t know the learning algorithms or the domain well enough, you cannot extract effective features & modify the algorithms for your own needs.

Just check the machine learning, data mining, information retrieval papers in the literature. All of those papers focus on improving the state-of-the-art techniques tailored for the specific task on a specific dataset.

Scott Brave

Keyser,

I agree with some of what you are saying. Yes, if you take a cross-domain, cross-application perspective, then very little can be done to generalize effectively.

However, when you limit scope to a domain (like ecommerce) and a set of applications (like customer segmentation, product recommendations, etc.), the data doesn’t necessarily look as different as you are suggesting. Sure, data is always somewhat different, and so you would need an expert to get the optimal benefit. But what I’m suggesting is that, with the right generalized approach, a domain expert (ecommerce expert in this example) could effectively work within a “friendly” system to derive significant benefit.

Mike Will

Nice. Thanks. Big data is still just data, though, not magic. ie Big Garbage In –> Big Garbage Out

Bryan M.

Exactly my thoughts.

Doesn’t really bring anything new to the table.

Shilpi Sharma

I do not think that data scientists are domain experts. Data Scientists and domain experts need to work together to build realistic models that are self-learning and constantly changing as and when business reality changes.

Jad9000

I agree. But, one of the issues with big data is that, in a perverse way, it sometimes is bundled in ways that make it too linear and a lot of the insights that could have been derived simply aren’t.

Alex

No. Expressing your real-world situation as a machine learning problem requires both domain specific knowledge and an understanding of machine learning. Here’s a simple example: you’re making a dating website, and you want to decide what people to suggest as matches. What is an instance? Is each historical pairing an instance? Is each person an instance? How do you select your labels? Is there a way to view this as a classification problem?

In practice, there is more to Machine Learning than taking a list of instances in a standardized format and applying black-box algorithms.

Yong Sheng

Obviously there are enough motivations and benefits to make big data easy to use. However, using the web content management tool as an analogy to big data is plausible: the quality of web content can be assessed by almost everyone; but the quality of big data product must be explained properly by a scientific mind. An untrained person is usually confused or misled by the metrics, and thus leads to costly wrong decisions.

Well, don’t get me wrong. I am not saying that data scientists are the smartest and others are stupid. The world of big data is like the situation of “blind men and the elephant”. We are all blind, and all the way we learn is to read the numbers. If what you read is the tail, you may think the elephant is like a rope. That’s the cruel nature of data science and thus requires scientific minds.

Robert Hawkins

There are two ideas that have been around for years and I think they apply here. One is that we are drowning in data, ie: producing data, information and reports that no one wants or reads. We are actually using more copy and printer paper than ever before. The other is that if you torture the data long enough, it will confess. People often take data or information and study and twist it to mean what they want while conveniently leaving out parts of it to make their case.

lucandrecormier

Comes to mind the David Ogilvy analogy about the drunen sailor who uses research (we now call it data) as a lamp post to lean against so as not to fall, rather than look ahead to move confidently forward.

Stephan Tual

Exactly Scott – while I’m a huge fan of Russel Jurney’s ‘agile data’ concept (http://www.slideshare.net/russell_jurney/hortonworks-roadshow), how many companies have the capacity/knowledge/finance to even consider hiring a team of Data Scientist to answer what are likely relatively ‘basic’ business intelligence questions?

What’s much more likely to happen is for a company’s existing Business Analysts to be trained up on a set of tools, such as Platfora, Datameer, Karmasphere or Pentaho’s Instaview, or even better – continue using the tools they know (SAS, R) through transparent connectors to Hadoop (RHADOOP). Odds are they will be able to deliver insight that in 99% of the cases will be ‘good enough’.

The 1% left likely has custom needs, domain-specific questions, requirements for specialized hardware/software and are likely to ‘roll their own’ regardless.

Scott Brave

True. But why not also package together some best practices on a per-domain basis?

Yin Cheng Lau

Interesting insights!

In addition to developing data scientists, we need to make analyzing data and deriving insights a way of life. It starts with whoever is in charge asking a simple question:

“Where is the data that supports what you propose and say?”

If all leaders from board members, senior management, to line manager persistently ask the above simple questions all the time, we will very quickly build a culture of analytics and fact-based decision making, similar to what great scientist have done in the past and still do. : )

Cory Peisel

Palantir Technologies. The solution is already out there, just not applied in this market.

Steve Ardire

Cory Palantir Technologies is impressive but please don’t shill or perhaps you’re just an enthusiastic forward deployed engineer ;)

Comments are closed.