It was possible to unite the two worlds, developers can write queries to generate the visualizations, but users also can do it with two more simplified interfaces for this, this created a very good interaction between the development team and the final users.There are other tools for data analysis like python with some libraries (pandas, matplotlib, seaborn, etc.), power bi, tableau, among others, but in this case the metabase was a great choice because: In our case, the main point was to find something that was quick to generate dashboards with metrics on the use of the system with a more managerial view, both for the heads of the areas and for the analysts. The metabase has a default for naming to make the table names more friendly, if you want to make your table names the way they are, you can change them at: Admin> General> FRIENDLY TABLE AND FIELD NAMES.If you add a table and it does not appear immediately and you know that everything is synchronized, perhaps it is hidden, if it is, you can change the visibility at: Admin> Data Model> (select the database)> (select the table) > VISIBILITY.It’s possible to save results of queries that take a long time to run, for this you need to enable it in: Admin> Settings> Caching. The metabase has an admin session, where it’s possible to register users and user groups, connect with new databases, set permission for collections and databases, in addition to a troubleshooting session. The visualization of the dashboard is very similar to that of X-Ray, the difference is that the dashboard is created by you with what you think is relevant, being possible to add static or dynamic filters. This will be your home page, if you have not connected to any database in the configuration step, the metabase provides a sample dataset. The metabase has the option to always synchronize, so as not to burden the application it may be better to connect the metabase to a read-only basis or disable this setting and synchronize only when necessary. SetupĪfter installation the configuration is very simple, you only need to create your user account, which will be used as an administrator, fill in the information to connect with the chosen database (remember to fill in the host and the port, even if it appears, because they are only suggestions), choose how your data will be synchronized and that’s it, you can start using the metabase. The JAVA_TOOL_OPTIONS parameter helps to limit the memory consumption of the application according to the memory consumption allowed for the container (more information here). In our case, we need to share the metabase information, so we chose to deploy using the database storage option, which can be used through the following docker-compose: version: '2' services: metabase-app: image: metabase/metabase ports: - "3000:3000" environment: - MB_DB_DBNAME=? - MB_DB_HOST=? - MB_DB_PASS=? - MB_DB_PORT=? - MB_DB_TYPE=? - MB_DB_USER=? - JAVA_TOOL_OPTIONS=? To avoid lose the data you have already configured (information from the metabase itself and not from your application, such as user registration, permissions, configuration of dashboards, etc.), if the container is turned off, it’s possible to point the metabase to a volume on your machine or use a database. docker run -d -p 3000:3000 -net=host -name metabase metabase/metabase To make a test on your local machine is very simple, you only need to execute the following command: docker run -d -p 3000:3000 -name metabase metabase/metabaseĪnd access: If the database you are going to connect to the metabase is at your localhost, you need to execute the same command, but use the net=host parameter, with that the metabase stops looking at the localhost inside the container and starts looking at the localhost of your host. There are several options for installing metabase, but we use it with docker.
0 Comments
Leave a Reply. |