Database Reference
In-Depth Information
Gutierrez:
What team do you work in and what is the makeup of it?
Heineike:
I work as part of the engineering team. We're building a Software-
as-a-Service platform to let our clients analyze text-based data sets at scale in
order to make important decisions. That means that are complex algorithms
and visualizations running, but they are wrapped inside of a software product
that clients interact with directly, and therefore has to work repeatedly under
lots of conditions. One of the cool things about working in an engineering
team is that there are people with skills that are very different from my own.
This means that it looks like they all have super powers. So between us, we are
able to build some things that are really interesting. We have UI developers
who build out the interactive visualization layer, we have back-end engineers
who are building the data collection, search, and analysis layer, a data science
team who work out how to work with the data and figure out improvements
to algorithms, and then QA and DevOps who make sure it's all running well.
One fascinating thing about Quid is that most data scientists work in compa-
nies where the main business of the company is something not related to data
science, whether it's finance or advertising or a social network or something
else. The data science used at these companies is there more or less to grease
the wheels and add interesting products and services on the side of the main
business. At Quid, everything we do is with data, so most of our team can
describe themselves as some form of data scientist. The people who work
here are generally very excited about this whole idea and curious about what
more we could be doing.
Gutierrez:
Tell me about Quid's product and how it uses network visualiza-
tion to help your customers understand the world.
Heineike:
For our users, the product first lets them search premium, large
volume data streams, and then presents the results in an interactive visualiza-
tion. We use network visualization as a way to organize the data to help peo-
ple explore it. We look at large, external, unstructured data sets that provide
signals about markets, consumers, and innovation, including news, patents and
data on startup companies. As an example, let's say you're analyzing the news,
in the visualization a node in the network might be an individual article. This
news article will then be linked to other articles that are very similar. When
you lay this out, you end up seeing clusters of things that are related to each
other, which lets you see the topics in the data. Then, between those clusters,
you will see these long range links where there are connections, but they're
not super close to each other, and so you can immediately get an idea of how
these topics relate to one another, and what bridges them. Then you can start
playing with the related metadata and laying that over the top, for example to
see the change through time. The product is very flexible from a customer's
point of view regarding the data that is visible and can be used. This allows
them to fully explore the space in a number of ways.
