How can I aggregate data from online sources about a specific topic?

It's unlikely that companies would look to outsource such a critical component and also it would be near impossible to create trust around 3rd parties accessing their data especially via an intermediary service.

First off, I have several people I could introduce. I'd also like to know the industry you're operating in, what the data looks like, where it comes from, and how much it needs to be cleaned up if putting into a relational database, or if the better solution would be a distributed file system like MongoDB where you don't necessarily need to normalize the data. Also, if you're a startup, or if the company is well established with many existing customers and if this is for a new initiative. Assuming you're working with a relational database, which it sounds like you are, you will want to implement something like Tableau or build out a custom dashboard using Google Charts, HighCharts, D3Js, or any number of other potential dashboarding/visualization solutions, which usually involves some programming/scripting in JavaScript. There are paid solutions like Tableau (which is amazingly powerful), and then there are free/open source options. I'd be happy to talk about possible ways to architect the solution, and discover who you would need once I understand the variables more. If you're building a web application, then you will likely need someone who is also a full stack developer, meaning they can handle building the back end and the front end in addition to the data requirements. Many early startups choose Ruby on Rails (because there is a ton of open source code out there for it) with Twitter Bootstrap (modified) and in order to visualize the data, they will need to work with JavaScript. It makes sense to have this person act as initial product and to derive the insights out of the data. They're pretty much the only person who can do this anyways, because they're the ones on the data. If you're in an early stage startup, I would recommend the strongest business owner (usually the early stage startup's CEO) be directly involved with this person on communicating what value your solution brings to clients, and what they pay you for, and in brainstorming on potential features and reports. Once the solution becomes established, and many customers start using it directly, there should be a different product person interfacing to those customers over time, gathering feature requests from customers and bringing it back to the Data Scientist/Analyst who spends their time working on the data. Depending on whether the solutions are SQL or NoSQL or hybrid, there are different types of Data Science professionals you should consider: 1. Data Scientist 2. Data Engineer 3. Data Modeler/Analyst 1. The Data Scientist handles experimenting with the data, and is able to prove statistically significant events and statistically valid arguments. Normally, this person would have modeling skills with Matlab, R, or perhaps SAS, and they should also have some programming/scripting skills with C++ or Python. It really depends on your whole environment and the flow of data. In my experience, Data Scientists that exclusively use SAS are sometimes extremely skilled PhD level statisticians and focused exclusively on the accuracy of the models (which is okay), but often not sufficiently skilled to fit within an early startup's big data environment in today's world and handle all of the responsibilities you'd like them to handle described in your question. I'm am not bad mouthing SAS people as they are often the MOST talented mathematicians and I have a great deal of respect for their minds, but if they do not have the programming skills, they become isolated within a group without a Data Engineer helping them along. Often a SAS user trying to fit into this environment will force you to use a stack of technologies that a skilled Data Architect would not recommend using. It takes programming in some object oriented language to fit into today's big data environments, and the better Data Scientists are using hybrid functional and OOP programming languages like Scala. Extremely hard to find Data Scientist can also work with graph databases like Neo4j, Titan, or Apache Giraph. 2. The Data Engineer, if you're dealing with a firehose of data like Twitter and capturing it into a NoSQL architecture, this is the person who would prepare the data for the Data Scientist to analyze. They often are capable of using machine learning libraries like WEKA to transform data, or techniques like MapReduce on Hadoop. 3. The Data Modeler/Analyst is someone who can use a tool like SAS, SPSS, Matlab, or even R, probably a very strong advanced Excel user, but likely won't be a strong programmer, although perhaps they will have a computer science degree and have some academic programming experience. The most important thing to watch out for is someone who is too academic, and has not been proven to deliver a solution in the real world. This will really screw you up if you're a startup, and could be the reason you fail. Often, the startup will run out of money due to the time it takes to deliver a complete solution or in the startup's case, a minimally viable product. Ask for examples of their work, and specifically dig into what it is that they did for that solution. I've tried to cover a pretty broad range of possibilities here, but it's best to talk in specifics. I'd be happy to discuss this with you in detail. To answer your question, is it perfectly reasonable for someone to handle all of the responsibilities described in your question, if you find the right type of person with the appropriate skills, and a history of success.

In my experience coaching hundreds of people, anytime you go against your gut (aka intuition) you will lose. Data is important to consider, but your gut often knows the path.

It's hard to get a definite answer to this as these companies will not tell us how their algorithms work. Having said that: Based on my experience working with these APIs, it is totally possible to implement a real-time search like "socialmention.com" purely based on the APIs from Google, Twitter, Yahoo, or Facebook. On the other hand, crawling and saving all that data just for the eventuality that a customer might search for it is probably not economically viable and will also violate the Terms of Services of most of these APIs. Bottom line is: They are probably not crawling and caching.

Qualifications on the answer: Created the first website to be ever commercially licensed by a ministry of education for in-school use (Brainium 1996). Was a VC associate at a 500,000,000 firm called Knowledge Universe focused on all things education-related that intersected with technology. Short Answer: Yes and no. Longer Answer: Learning analytics is one of those "lightening in a bottle" kind of industries. Everyone knows it's going to happen, whomever makes it happen at scale first, will have a huge advantage, and even more niche plays in this space will still create massive value for shareholders and significant transformation of education as we know it. That's the good news. The problem is that funding startups in the education space that are dependent on "permission" being granted by existing institutions and their employees is very much hit or miss. It's getting easier to raise a $500-800k seed round for a good idea with a good team, but the problem is that getting the next bigger round is very challenging in that the models of most of these seed-funded companies are not growing fast enough to prove to institutional investors that a growth model of the idea being very big in a relatively short time-period is credible. So if you're intent on pursuing this, I'd be sure to focus entirely on how you can get growth and adoption with requiring little to no institutional buy-in which sounds almost impossible on the surface of the area you're exploring but I believe isn't actually impossible. I'm very passionate about this space and want something like what you're describing to succeed so would be happy to do a quick call to hear where you're at and see if I can provide you some actionable ideas on how to reduce the friction / improve adoption of your analytics product.