Log all actions on a website to Data Foundry, including scroll behavior, mouse movement and browser characteristics.
For digital design, there are numerous methods to do UX research on the interaction flow of the UI (e.g. website, app or dashboard). One of them could be customer journey logging. You can program your UI in such a way it tracks the actions from a visitor from that specific website (e.g. scroll behaviour, mouse movement, browser characteristics), in order to find out what an optimal user flow might in order to achieve satisfaction of the user, or in the context of webshops, can be used as a marketing tool.
Only collecting this data might not be enough to find valuable patterns or discover new relations in the data. Therefore, using the Data Tool in DF makes is easy to visualize your data streams and apply sampling where needed. Next to that, data mining and AI can be applied when you want to apply supervised and unsupervised learning techniques to your data, in order to find correlations.
Customer Journey Logging is a method often used to analyze the customer flow of webshops. By being able to collect data on what type of pages with specific projects are spending their time on and for how long, companies can use customer journey logging to get a profile of what kind of products a specific person likes, and how they can use that data to recommend you other products that are in line with your taste.
A project created in Data Foundry and inside the project an new IoT dataset with
openParticipation switched on (edit project, see bottom of form). The IoT dataset needs to be configured to get input data from an OOCSI stream (dataset page, see second configuration tab). The channel that you have configured in the OOCSI stream of the dataset, needs to be inserted also in the code below.
Use the OOCSI network to add data items for the IoT dataset. This is option is available from a wide variety of platforms and technologies. A list with possible options is provided under the tab OOCSI API.
In the part that starts with
OOCSI.send("CHANNEL NAME",, you first need to change the CHANNEL NAME to the one provided in the Data Foundry configuration of the IoT dataset. Then you can determine what type of web tracking data you want to log into the IoT dataset. Right now it only focuses on mouse position when the visitor clicks somewhere on the page. This can be further extended, for instance, with browser characteristics, time that a visitor spends on a single page, user interaction flow, etc. You can change these things in this line to make it send the right type of data to Data Foundry.
There are two ways to export your data. If you want a simple export without any preprocessing, visualization or annotation, you can use the download button in the dataset screen. If you want a more advanced export, you can use the Data Tool to export your data.
For more info on exporting with the Data Tool, click here for a basic explanation.
A useful next step for this application is using data mining and machine learning techniques in order to make actual sense of the data. There are various tools and platforms for doing this, that can connect to Data Foundry to retrieve the data locally. One example is already given in the platform: if you go to a dataset page and open the CSV token link tab under configuration, you will find a code example in Python to download and visualise the dataset contents. You can use similar code to work with the dataset in Python for machine-learning purposes. This will only work if you generate a token link first.