This tutorial is at a preliminary stage. Please send your feedbacks (concrete comments and possible solutions) to Issues.

Although you can anonymously browse datasets and visualisations, more functionalities will be available after logging in. Two types of accounts are accepted by this site, a Web Observatory account (wo account), or a Southampton University account (su account). An wo account can be created at the signup page.

You can login using either account at the sign in page, by choosing the desired login button. For wo account login, the username is your email. For su account login the username is your university username (without part), and the password is your university password (not ecs password). The first time you login using your su account, a wo account will be automatically created. This account will use your university email as the username, and a password reset is required the first time you use it.

Datasets and visualisations registered on this site can be browsed at either Datasets and Visualisations respectively. The default view shows a table of entries with basic information (metadata). You can specify how many entries are shown per page, filter entires using the search box, and reorder entires by clicking the title of a column. A detailed information panel along with additional buttons will show after clicking the name of an entry. Function of these buttons will be described in section Query datasets.

Each entry on this site has a uniques identifier (e.g. DBpedia /wo/dataset#52e19220bef627683c79c3a6). These identifiers are resolvable and will give detailed metadata of the entry, as if you've clicked the name of the entry. The segment part (i.e. the string after #) is the ID of this entry, which is required to used the Query API.

You are more than welcome to register your datasets and visualisations on this site, which can be doen by clicking the Add button at either Datasets or Visualisations. You will have to provide basic metadata, such as name, URL, publisher (you), creator etc. as well as access control options at the bottom. For datasets, you have to provide username and password of datasets where applicable. In addition you need to test whether the URL, username and password you provided are valid.

Once you logged in, you can query a dataset that you have permission to access (i.e. either publich datasets or the publisher has granted you permission). An Explore button will become available on top of the detail panel of the dataset. Clicking that button will show a query panel below the detail panel. Currently supported datasets/query languages are SPARQL, MongoDB and MySQL.

The query API allows users to create applications that access/query datasets on users' behalf. To enforce protection for (private) datasets on the portal, the query API is protected by OAuth2.0, the technique that enables you to login to various sites using your Facebook/Google account. To use the query API firstly you need to follow the following steps:

  1. Create a new application at Applications in your profile, where you can have an applicatoin ID and a secret. They're the credentials of your application which should be kept safe.
  2. Authenticate and authorise your application through OAuth 2.0. Here we provide an example using the Resource Owner Password Credentials Grant. In the example capitalised words represent variables to which a user should provide values. Currently the API can only be used by a

    HTTP POST to with the following data fields grant_type:'password' client_id:APP_ID client_secret:APP_SECRET username:USER_NAME password:PASSWORD

    jQuery example :

         {grant_type:'password', client_id:'53be91cb9790d31811850792', client_secret:'dien8ck4kdie3', username:USER_NAME, password:PASSWORD},

    httpie example :

     http POST grant_type="password" client_id="53be91cb9790d31811850792" client_secret="dien8ck4kdie3" username=USER_NAME password=PASSWORD

    cURL example :

     curl --data "grant_type=password&client_id=53be91cb9790d31811850792&client_secret=dien8ck4kdie3&username=USER_NAME&password=PASSWORD"

    Note: To access https using httpie or cURL you have to configure them first.

    The response from the server will contain an access token if the request went through.

    Example response:

     { access_token: "BAd8AFoQI6ydBWF59d48WXrCpAEqiQG1Mn5ai3A3UEE=",
       refresh_token: "woHiAGTspWnsYCJD0ACv9yOi4sMxZln0lWix99dd4hU=",
       expires_in: 3600,
       token_type: "Bearer" }
  3. Use the access token to query a dataset. HTTP GET (DATASET_ID can be obtained from the URL of detial information page, the string after #), with the data field query:QUERY_STRING, collection:COLLECTION_NAME (different datasets may require different fileds, which can be found from the query page of a dataset) and the header Authorization:"Bearer ACCESS_TOKEN"

    The example below queries the Ebola Tweets dataset whose ID is 548047c832582f2c45520ce7.

    jQuery example :

         type: 'get',
         url: '',
         data: { query:'{}', collection: 'tweet' },
         headers: {Authorization:'Bearer BAd8AFoQI6ydBWF59d48WXrCpAEqiQG1Mn5ai3A3UEE='}
         }).done(function(result) {

    httpie example :

     http GET Authorization:"Bearer BAd8AFoQI6ydBWF59d48WXrCpAEqiQG1Mn5ai3A3UEE=" query=="{}" collection=="tweet"

    cURL example :

     curl  --header "Authorization:Bearer BAd8AFoQI6ydBWF59d48WXrCpAEqiQG1Mn5ai3A3UEE=" "{}&collection=tweet"

    The response from the server will contain the solution of your query.