One Search – presentation transcript

Hi!
My name is Niklas, and I work as a system specialist at KTH Royal Institute of Technology here in Stockholm. For those of you who doesn’t know about KTH we are one of the biggest technical universities in northern europe and we account for one-third of technology education and science in Sweden. We have about 13000 students and 4500 employees and scientists.

Today I’m going to talk about our search project that we are working on, and the vision of “One Search” – One natural place for employees and students to search and find information that they are looking for.

The team for this project consists of one web coordinator, three in house developers and one consultant, Niklas Lind, who is here today from Valtech.

Agenda

So, this is the agenda. We’ll start with the challenges that we had in the beginning of the project with a large organization, a big website and a bad search experience.

Then I’ll talk about what we have done so far in the search project and the concept of “One Search”, the addition of business knowledge and secure data.

And at the end we take a look on the results and how we see the future of search at the university.

Challenges

The challenges that we faced in the beginning of the project was that we had several different search function connected to different data sources. This lead to the student or scientist needing to know which search function to use depending on what information they’re looking for.

We didn’t have any business knowledge added to the search function and no secure data which mean that a lot of important information was not searchable.

But there were two things that we did have; 20,000 employees and students. Which means complaints and a big messy search index.

So at least we knew where to start!

What have we done so far?

First of all we installed the Google Search Appliance and let it crawl all of our data sources. We then realized that we had 2,000,000 webpages and about 650 sub domains. So the first action point was to clean up the index so we knew what we had to work with.

The next thing was adding business knowledge. And when I say business knowledge I mean building entities like courses, programs rooms, people, etc. It also includes relations between these entities, so we can view the information in different aspects. Let me take an example:

Example:
Let’s say a student comes up from the subway at KTH and have an exam today. He doesn’t know when or where the exam is so he picks up his phone and enters the course code or course name in the search box and hits the search button.

The first result is the course result where he can see when the exam is and in which room. He clicks the link for the room and get’s a detailed description of the room with an integrated google map and a geo guidance from his current location to the room where the exam is going to be held.

Another example, let’s say you are looking for a person and you don’t know her last name. You search for ”Lisa” and in the result you find fifteen Lisa. But because Lisa has added an image of her self in her profile, and that image is shown in the result result, you can easily determine which Lisa you are looking for. We have also added contact information like e-mail and phone number in the result so you don’t have to click further down to her profile to be able to contact her.

Then we started working with personal and secure data. At KTH we have a quite big social intranet similar to Facebook, but with a learning focus, where students and teachers communicate. This information is often personal and secure, so we feed that to the Google Search Appliance and make it searchable.

At the end we modified the design to be able to deliver enriched search results and analyse the search. We use Google Analytics to analyze what visitors are searching for, which search terms they use and where they click in the list of results. This way we can continue improving the search solution.

All this, together with all the data sources indexed in one search engine, added business knowledge and secure data with an integrated design is our vision of the “One search”.

Results and Future

The results so far is a highly improved search experience where the users can find the information they want, when they need it.

By using the Google Search Appliance we reduce maintenance costs by not having to build admin interfaces, come up with ranking algorithms and so on. Instead we can focus on the business values, which are more important for us.

At the end we make more information searchable and easier to find, which leads to an increase of knowledge exchange. Information is everything and a key factor for KTH to stay on top in the future.

In my opinion we already have the best search in Swedish education, but this is a continuous work and we have just started.

Lämna ett svar

E-postadressen publiceras inte. Obligatoriska fält är märkta *