
Yes, you can download the complete CSV file of the processed tables directly on the platform.
However, tables with more than 200,000 rows can only be accessed through our data lake on BigQuery or our Python and R packages.



Access to most of the data offered by DB is free. Payment may be required in only two cases:

1. Some data is part of our advanced plans, Pro and Orgs, as indicated on the table pages. In this case, a subscription is required to access the data.

2. One way to access our data is through BigQuery, on Google Cloud, which offers up to 1 free Terabyte per month for its users. We note that this limit is usually sufficient, even for analyses using larger datasets, such as RAIS or the School Census. If you exceed this limit, BigQuery charges 5 dollars per terabyte of data your query processes.



You can access the LAI requests by clicking the redirect button that links out to the external platform.



You can access the original data sources by clicking the redirect button that links out to the external platform.



You can query DB's processed tables in different ways. Besides downloading directly on the platform, you can explore the data with SQL through BigQuery or with the Python and R packages.



Our packages allow direct access to the public *datalake* from your computer or development environment. To start exploring our data in Python or R, follow the package installation tutorials in our [documentation](https://basedosdados.org/docs/access_data_packages/).



The vast majority of the datasets found on our platform are public. However, some private datasets are also listed. Our organization is nonprofit, so we do not charge for listing datasets on our platform. We believe it is useful to list private datasets, since people may be interested in acquiring them after finding them on our site.



Some data is updated automatically and some is not. We are committed to automatically updating any dataset with monthly or higher frequency. This data is available on our advanced plans, Pro and Orgs.



These are tables that link various institutional codes and information from different Brazilian entities. This matters because it solves the problem of there being no unique identifier for entities — such as state, municipality, school, district, census tract, ICD-10 and ICD-9 categories, CBO-2002 and CBO-1992, etc. — across institutions, or of identifiers and names changing, with typos, across years and institutions. To learn more, check out our [tutorial](https://medium.com/basedosdados/diret%C3%B3rios-brasileiros-como-essa-base-facilita-sua-an%C3%A1lise-40dc8ce2ca2) on Brazilian Directories.



These are requests made under the Access to Information Law (LAI).

The LAI, Law No. 12,527/2011, regulates the constitutional right of any person to request and receive public information produced or held by public bodies and entities, across all levels of government and branches of power.



These are links to pages outside the platform with useful information about the dataset. Most of these are data we have not yet made available on DB as processed tables. We always try to provide the closest possible path to the source to download the original data.



These are tables ready for analysis, available via SQL, Python, and R. The process of preparing the tables involves standardizing variable names and making codes compatible, which makes it as simple to cross-reference tables from different institutions and topics as it is to run any other query.



The download limit for data on our platform is 200,000 rows.
To access tables that exceed this limit, use our *datalake* on BigQuery or our Python and R packages.





The level of observation indicates what each row of the table represents. It is as if the combination of columns that make up the level of observation were a primary key. This information is useful because it indicates the smallest possible granularity of analysis with that data. For example, a table with a state-level observation allows analysis at the country level (broader than state), but not at the municipality level (a more specific breakdown).



On our platform, you can find three types of data: treated tables, original sources, and LAI requests.



It is very important for us to know which data our community needs, so the team can add the requested datasets to the data prioritization process. We have a Discord channel set up to receive your proposals and suggestions — just follow the steps indicated there. Access it [here](https://discord.gg/tx57ek6zqQ).



All data available on DB Pro is updated automatically through our flow management technology. DB uses state-of-the-art data engineering to guarantee the highest quality and consistency in what we provide.



We currently have two plans. The Pro plan costs R$47 per month, giving a single user access to all up-to-date data. The Orgs plan costs R$385 per month, including access for 10 accounts and priority support from our team.



In addition to the free plan, DB offers advanced plans: Pro, for individual users, and Orgs, for up to 10 users from the same organization. These plans provide, on top of the data already available in DB's public data lake, dozens of exclusive datasets prioritized to solve problems and help you, your team, your organization, or your research. Besides exclusive datasets, the plans offer access to the most up-to-date version of some of our key tables, such as CNES, CNPJ data, and Brasileirão data, among many others.



The subscription grants access to up-to-date data. This can sometimes be a subset of a larger processed table, or an entire processed table. Our search engine and specific dataset pages indicate what is available for free and what is covered by the paid plan subscriptions.



To create a project on Google Cloud, you just need an email registered with Google. You need to have your own project, even if empty, to run queries on our public *datalake*.

1. Go to [Google Cloud](https://console.cloud.google.com/projectselector2/home/dashboard). If it's your first time, accept the Terms of Service.
2. Click *Create Project*. Choose a good name for the project.
3. Click *Create*.



Partitions are divisions made in a table to make data management and querying easier. By segmenting a large table into smaller partitions, the number of bytes read is reduced, which helps control costs and improves query performance.



To check a sample of all the variables in the table, use:


```sql
SELECT * FROM dataset.table_name LIMIT 100
```

- The first valuable tip is to select only the columns you are going to use. BigQuery works using a columnar model, so the fewer columns you use, the better your query's performance will be. This means avoiding the classic

  ```sql
  SELECT * FROM table_name
  ```
and choosing the columns you are interested in instead. It sounds tedious, but it helps a lot!

- For large tables, a good practice is to filter the years and states you are interested in with the

  ```sql
  WHERE
  ```

clause. Since we use a partitioning system, this will drastically reduce processing cost and time.

These are the simplest and quickest tips to apply. To learn more about best practices, check out the [complete manual](https://cloud.google.com/bigquery/docs/best-practices-performance-overview?hl=pt-br) provided by Google Cloud.



These are the data types of Google Standard SQL. Their categories include: INTEGER, STRING, DATE, FLOAT64 (Decimal), GEOGRAPHY, among others.



If you exceed the limit, you are charged 5 dollars per Terabyte of data your query processes. However, we note that Google's 1 Terabyte limit is usually sufficient, even for analyses using larger datasets, such as RAIS or the School Census.



It is a public database hosted on BigQuery, Google Cloud's powerful data storage and analysis tool, with all of DB's treated tables for you to query directly from your browser. The advantages of using our *datalake* on BigQuery relate to processing speed, scalability, and cost savings.



Our public database is hosted on BigQuery, a powerful Google Cloud data storage and analysis tool. For this service, Google Cloud offers up to 1 Terabyte free per month to its users. You therefore need to create an account to link the database's queries to you and track the total processing used in the month. Creating the account requires no card or prior payment method, since BigQuery starts automatically in Sandbox mode, which lets you use the resources without adding a payment method.



DB Lab was created to ensure the organization's sustainability, so as to maintain and expand our products free of charge. Providing data services and carrying out projects with partners has been key not only to generating revenue, but also to strengthening our position as a reference in making quality data available.



DB Lab is our services, partnerships, and projects arm. It is through DB Lab that we interact with other organizations, providing our accumulated expertise in technology and data. We see these projects as an indispensable component of our mission to promote the use of data for social well-being.



Yes, whenever a course is completed we issue an official DB certificate with the student's personal information and course details. The certificate is verifiable and can be shared on social media or attached to your résumé.



Yes, you can send your questions to [suporte.bdedu@basedosdados.org](mailto:suporte.bdedu@basedosdados.org). Our team checks this inbox daily to provide the best support possible.



We currently offer data analysis courses covering tools that matter in the job market, such as SQL, Python, spreadsheet editors, and Google Cloud. Our team is also equipped to develop courses on demand, and we are working to launch new training formats.



DB Edu is our education arm, with the mission of empowering people to solve problems with data. We have provided training services to organizations such as Vetor Brasil, Souk Analytics, and ProForest, in addition to our courses open to the public. We offer live and asynchronous courses, and all proceeds are reinvested in Data Basis.



We use the coda.io platform, which works as a document editor combining features of spreadsheets, presentations, word processors, and apps. There you'll find our Student Area, where you can access the schedule, links to live classes, recorded classes, an exercise submission page, a Q&A forum, extra materials, and more.



Our courses are developed and reviewed by specialists who work on the Data Basis team. You can check everyone's profile on our [institutional page](https://basedosdados.org/about-us).



You can reference us in two ways:

- Just with the name "Base dos Dados" (in Portuguese), "Data Basis" (in English), or "Base de los Datos" (in Spanish);

- From the *white paper* [Data Basis: Universalizing Access to High-Quality Data](https://osf.io/preprints/socarxiv/r76yg).

> Dahis et al. (2022) Data Basis: Universalizing Access to High-Quality Data. Available at: {`<osf.io/preprints/socarxiv/r76yg>`}



We collect personal information that you provide directly to us when using our services. This may include your name, email address, payment information, and other information you choose to share with us. We also collect information automatically, through cookies, including usage data, IP address, browser type, and information about the device you use to access our services.

We use your personal information to provide and improve our services, including processing payments, providing customer support, and personalizing the user experience. We will not sell, rent, or share your personal information with unaffiliated third parties without your explicit consent. For more information, see our Terms of Use and Privacy Policy [here](http://basedosdados.org/terms).



Over the years, DB has grown with the financial support of users and people who identify with our mission. Today, we also rely on our Services and Partnerships area (DB Lab) and our plans for access to exclusive, updated data (DB Pro). Learn more about the organization's accounting on our [Transparency](https://basedosdados.org/transparency) page.



No. DB is constituted as a nonprofit association called "Instituto Base dos Dados," headquartered in Rio de Janeiro, with company registration number (CNPJ) 42.494.318/0001-16.
