replace superset with evidence

Author: sdebruyn

.evidence/customization/custom-formatting.json

@@ -0,0 +1,4 @@
+{
+  "version": "1.0",
+  "customFormats": []
+}

.gitignore

@@ -89,3 +89,8 @@ $RECYCLE.BIN/
 
 assets/*
 !assets/.gitkeep
+*.duckdb
+.evidence/template
+.svelte-kit
+build
+node_modules

README.md

@@ -1,6 +1,6 @@
 # in⚡️ight
 
-A simple data project using [dbt](https://getdbt.com), [DuckDB](https://duckdb.org/) and [Superset](https://superset.apache.org/) to analyse your electricity data from Belgian smart meters.
+A simple data project using [dbt](https://getdbt.com), [DuckDB](https://duckdb.org/) and [Evidence](https://evidence.dev/) to analyse your electricity data from Belgian smart meters.
 
 Thanks to [Jacob Matson](https://github.com/matsonj) for [his blog post](https://www.dataduel.co/modern-data-stack-in-a-box-with-duckdb/) that inspired all of this.
 
@@ -12,14 +12,18 @@ The dbt documentation is available at [https://sdebruyn.github.io/inzight](https
 
 ### Requirements
 
-* Python 3.8 or newer
-* Docker (for Superset)
+* Python 3.8 or newer (dbt & DuckDB)
+* NPM 7 or newer and Node.JS 14 or newer (Evidence)
 * Your source data (see below)
 
 ### Setup
 
-1. Clone with all submodules: `git clone --recurse-submodules https://github.com/sdebruyn/inzight.git`
-1. Install the dependencies: `pip install -r requirements.txt`
+1. Clone the repo: `git clone https://github.com/sdebruyn/inzight.git`
+1. Install the dependencies:
+    ```bash
+    pip install -r requirements.txt
+    npm install
+    ```
 1. Create a profile named `inzight` in `~/.dbt/profiles.yml` with the following content:
 
     ```yaml
@@ -29,8 +33,7 @@ The dbt documentation is available at [https://sdebruyn.github.io/inzight](https
         dev:
           type: duckdb
           threads: 12 # should be the number of cores in your system (or double if you have hyperthreading)
-          database: /tmp/inzight.duckdb # or any other path on your system
-          external_root: /Users/Quack/code/inzight/assets # point to the assets folder in this repo
+          database: /path_to_the_project/inzight.duckdb # point to the project folder
     ```
 
 ### Source data
@@ -44,16 +47,7 @@ My column names are in Dutch, I have no idea what you get if you have a differen
 1. Make sure your source data has been added as a file named `assets/fluvius.csv`
 1. Run `dbt deps` to install the dbt package dependencies
 1. Run `dbt build` to create all models
-1. Run `docker compose up -d` to start Superset
-1. Open `http://localhost:8088` in your browser to go to Superset
-1. Log in with the default credentials: `admin` / `admin`
-1. Create a new database (*Settings > Database Connections*) with the following settings:
-
-    | Setting | Value |
-    | --- | --- |
-    | Display Name | `inzight` |
-    | SQLAlchemy URI | `duckdb:///:memory:` |
-1. Register the data marts in Superset as the following query: `select * from '/app/assets/name_of_the_mart.parquet';`
+1. Run `npm run dev` to start the Evidence server and open the dashboards in your browser
 
 ## Note
 
@@ -62,5 +56,3 @@ It's (like most hobby projects) a never-ending work in progress. Contributions a
 ## License
 
 MIT License
-
-Code under `superset/` is licensed under the Apache License 2.0 and taken from the [Superset repository](https://github.com/apache/superset)

dbt_project.yml

@@ -15,4 +15,4 @@ dispatch:
 models:
   inzight:
     marts:
-      +materialized: external
+      +materialized: table

docker-compose.yml

@@ -42,6 +42,22 @@ models:
         description: Average of the last 12 peak values including this month (used for your capacity tariff)
         tests:
           - not_null
+      - name: month_name_short
+        description: Short name of the month
+        tests:
+          - not_null
+      - name: month_name
+        description: Name of the month
+        tests:
+          - not_null
+      - name: month_start_date
+        description: Start date of the month
+        tests:
+          - not_null
+      - name: pct_change
+        description: Percentage change of the peak value compared to previous month
+        tests:
+          - dbt_utils.at_least_one
 
   - name: mrt_validated
     description: |

models/marts/marts.yml

@@ -7,11 +7,14 @@ with usage_data as (
 
 components as (
     select
-        month(from_timestamp) as month,
-        year(from_timestamp) as year,
-        from_timestamp,
-        usage * 4 as quarter_power
-    from usage_data
+        month(u.from_timestamp) as month,
+        year(u.from_timestamp) as year,
+        u.from_timestamp,
+        u.usage * 4 as quarter_power,
+        dd.*
+    from usage_data u
+    left join {{ ref('dim_dates') }} dd
+    on u.from_timestamp::date = dd.date_day
 ),
 
 month_peaks as (
@@ -23,19 +26,29 @@ month_peaks as (
     group by month, year
 ),
 
+with_avg as (
+    select
+        d.month as month,
+        d.year as year,
+        d.month_name as month_name,
+        d.month_name_short as month_name_short,
+        d.month_start_date as month_start_date,
+        d.from_timestamp as month_peak_timestamp,
+        d.from_timestamp::date as month_peak_date,
+        p.month_peak as month_peak_value,
+        avg(p.month_peak) over (rows between 11 preceding and current row) as month_peak_12month_avg,
+    from components d
+    join month_peaks p
+    on d.month = p.month
+    and d.year = p.year
+    and d.quarter_power = p.month_peak
+),
+
 final as (
     select
-        components.month as month,
-        components.year as year,
-        components.from_timestamp as month_peak_timestamp,
-        components.from_timestamp::date as month_peak_date,
-        month_peaks.month_peak as month_peak_value,
-        avg(month_peaks.month_peak) over (rows between 11 preceding and current row) as month_peak_12month_avg
-    from components
-    join month_peaks
-    on components.month = month_peaks.month
-    and components.year = month_peaks.year
-    and components.quarter_power = month_peaks.month_peak
+        *,
+        (month_peak_value - lag(month_peak_value) over (order by month_start_date)) / lag(month_peak_value) over (order by month_start_date) as pct_change
+    from with_avg
 )
 
 select *

models/marts/mrt_capacity_tariff.sql

@@ -6,6 +6,6 @@ Welcome to the auto-generated documentation for [In⚡️ight](https://github.co
 
 #### Available datasets
 
-The available datasets can be found in the marts folder in the project. These are materialized as parquet files in the assets folder.
+The available datasets can be found in the marts folder in the project.
 
 {% enddocs %}