Save your MotherDuck token there, and you'll see some charts already there.
This example uses a MotherDuck share, which is available by default for all users under the `sample_database` database.
For our dashboard, we'll use a dedicated share called `duckdb_stats`.
### Building a PyPI dashboard
Now that we understand how the basic template works and we are connected to MotherDuck, let's create some basic charts based on the data on the `duckdb` Python package.
I'll walk you through 2 examples : A big value component and a line chart.
Note that Evidence provides much more charts component on their [documentation website](https://docs.evidence.dev/components/all-components/).
#### Defining the source query
We'll define a source query that will feed our charts.
This needs to be located under `sources/motherduck/`. We'll compute weekly downloads as the source table gives us daily downloads but we don't need that level of granularity, so we'll aggregated it to fetch less data and have better performance in general on our dashboard.
The file name will be the table name that we point to in our `.md` file.
Here's the content of `sources/motherduck/weekly_download.sql`
```
SELECT
DATE_TRUNC('week', download_date) AS week_start_date,
version,
country_code,
python_version,
SUM(daily_download_sum) AS weekly_download_sum
FROM
duckdb_stats.main.pypi_daily_stats
GROUP BY
ALL
ORDER BY
week_start_date
```
Few comments :
* The `DATE_TRUNC` function simplifies handling dates by returning the start date of the week. This makes it easier to read and understand than just a week number. Additionally, it allows for straightforward grouping by week.
* `GROUP BY ALL` is a handy feature in DuckDB that makes SQL queries easier. You don't have to list each column you want to group by—this feature does it for you. It also makes your queries easier to change later. Other databases [have started to implement this function too.](https://www.linkedin.com/posts/mehd-io_sql-activity-7168265860292280320-F_0l?utm_source=share&utm_medium=member_desktop)
#### Adding the charts in the markdown file
Building a chart is a 2 steps processes, everything in Markdown.:
* Defining the SQL query
* Using the chart component
We'll have one page, the markdown file is located at `/pages/index.md`.
**Big Value chart**
The query will look like this :
Each query is wrapped with the `sql 
### Line chart
We want a weekly download view for the line chart, so it's simple as a group by on the `week_start_date` we calculated from our source query
```
SELECT
week_start_date,
SUM(weekly_download_sum) AS weekly_downloads
FROM
weekly_download
GROUP BY
week_start_date
ORDER BY
week_start_date DESC
```
The component is then :
```