& UX/UI Design
Location: Santiago, Chile
Skills & Tools
Marcelo Surjan, Manuel García, Begoña Taladriz
Graiph is a software that uses Machine Learning to standardize the drill holes logging process in geology, in order to improve data quality. The app seeks the implementation of technological tools that use Artificial Intelligence to complete the task in less time, with automation functions and with the possibility of easily accessing to the information of each drilling project.The copper mining sector in Chile is one of the pillars of the Chilean economy. Despite this, the exploration of this mineral has always been a difficult job, in which geologists play a key role, and the most difficult part of the process is logging drill holes. It's a tedious and slow task, which often delivers inconsistent and inaccurate results.
To begin, we define a 6-steps work method: (1) Discovery - current interaction analysis, (2) On-field Observation, (3) Interaction Design, (4) Product Design and (5) Testing & Redesign.
Step 1: Discovery
The starting point: current interaction analysis
A key objective of this phase was to understand the reasons why users (and companies) continued to prefer using the traditional tools and/or methods to implement the drill-holes logging, that is, to do it manually. This method takes a geologist approximately 12 hours to register 100 meters of drill-core. A drill-core can have up to 1,800 meters in total.
We realized that, although there were some digital tools, such as GVMapper and others, geologists preferred to use traditional techniques because those tools require some type of additional equipment to perform the registration correctly, as well as a large investment of the company.
After clearly seeing the context and the problem, there was a need to dimension who was the person involved in that scene. To understand more deeply the user we were going to work with, we created a persona who represented him/her, taking the first research as a basis. We built empathy and journey maps to organize the key interactions of this early problem we were putting together.
Once we pictured our theory, we start organizing the next step: On-field observation. For that, we set early-goals to test while on-field, in order to allow an agile development of the project:
Reduce the interaction time to 70% in a first stage.
Allow and optimize the access of the remote teams to the data generated by the geologists and the app.
Standardize the format of the data delivered.
Enable secure data storage.
Step 2: On-field Observation
Minera Las Cenizas
We traveled to Minera Las Cenizas, a company that was interested in collaborating with the project. There, we observed how 2 geologists worked on the interaction under study. In addition to verifying the objectives that we had set, we reached the following conclusions:
The data record is fragile and poorly standardized.
The geological analysis depends a lot on the geologist himself. Two geologists can reach very different conclusions from the same drill-core.
There are tasks within the process of logging that are easily automatable.
Connecting to a network to upload data can be a problem in the future that should be considered.
Possible ambiguities in the recognition of type of mineral: what kind of data should feed the software?
It was a key decision the one about standardize the software input: photo, video, illustration, interaction or other.
Step 3: Interaction Design
The first approach to the new app. we developed was a new user flow. We were aware that usability was key to achieving a good rate of effective users in the future. In that sense, we begin the process by defining a minimal linear interaction, from which to start:
Home: enter a new sample.
The photo is taken, synchronized or downloaded to the device.
The photo is uploaded to the app. (input).
The photo is processed.
The software delivers results: proposal of analysis of elements (geotechnics, lithology, alteration) as multiple layers.
The user verifies the 3 outputs (can see layers simultaneously or separately).
The user corrects the analysis proposal (add, delete or modify the data).
Once corrected, the current analysis is uploaded to the cloud.
The user and the team can view the data online and can download it as spreadsheets.
After verifying with some users the main linear interaction defined by the team, the first user flow was designed, as shown in the following image:
Step 4: Product Design
Wireframes and alpha prototype
We developed the first wireframe for the main interaction, in order to test it as soon as possible. We decided that the app. must allow 4 actions directly: enter a new analysis, review saved (unpublished) analyzes, view all team analyzes (published) and view team activity. Therefore, we design a permanent bar at the bottom of the interface to allow shortcuts to each of these actions.
Style and visuals
For this phase, we created a new brand and design system, so that the developer and the rest of the team would align at the time of creating the code. We chose bright colors and a clean typeface that works well in small sizes, because we needed to add lists with large amounts of data, all in compact interfaces.
Step 5: Testing & Re-Design
In collaboration with the team, a field test with functional prototypes and the new visuals was scheduled. It was a full day at Minera Las Cenizas, testing with tablets and receiving feedback from geologists. After that, we made changes in the design and code, most associated with geological interpretation variables.
Once the first redesign was ready, the app was finally developed and launched as a real prototype at Minera las Cenizas. Since that time, we have been constantly improving the interaction design with the feedback of the real users.
Thanks to the design and development of the software the company went from making consulting services to provide technological solutions.
In terms of improvement, the app. allows users take 50% less time to do the task of logging.
One year after launching the new UX/UI, the sales have increased to 700%.
Graiph makes 500% the conversions the other similar services achieve.
We are currently starting an implementation for AMSA, a 4.733 billion USD revenue company in Chile.
We are also in a project with IDIEM, a Research and Development Center under the Faculty of Engineering of the University of Chile.