chore: Add SQL scripts for revenue and delivery analysis

sql/delivery_date_difference.sql

@@ -0,0 +1,29 @@
+-- Calculates the average difference in days between the estimated delivery date and the actual delivery date for all orders that have been delivered.
+--
+-- Explanation step by step:
+-- 1. Select the customer state and the average difference in days between the estimated delivery date and the actual delivery date for all orders that have been delivered.
+-- 2. Join the olist_orders table with the olist_customers table on the customer_id column.
+-- 3. Filter the results to only include orders that have been delivered and have an actual delivery date.
+-- 4. Group the results by the customer state.
+-- 5. Order the results by the average difference in days between the estimated delivery date and the actual delivery date.
+SELECT
+    oc.customer_state AS State,
+    CAST(
+        AVG(
+            julianday (
+                STRFTIME ('%Y-%m-%d', oo.order_estimated_delivery_date)
+            ) - julianday (
+                STRFTIME ('%Y-%m-%d', oo.order_delivered_customer_date)
+            )
+        ) AS INTEGER
+    ) AS Delivery_Difference
+FROM
+    olist_orders oo
+    JOIN olist_customers oc ON oo.customer_id = oc.customer_id
+WHERE
+    oo.order_status = 'delivered'
+    AND oo.order_delivered_customer_date IS NOT NULL
+GROUP BY
+    oc.customer_state
+ORDER BY
+    Delivery_Difference ASC;

sql/get_freight_value_weight_relationship.sql

@@ -0,0 +1,23 @@
+-- Calculates the total freight value and the total weight of the products in each order where the order status is 'delivered'.
+--
+-- Explanation step by step:
+-- 1. Select the order ID, the total freight value, and the total weight of the products in each order where the order status is 'delivered'.
+-- 2. Join the olist_orders table with the olist_order_items table on the order_id column.
+-- 3. Join the olist_order_items table with the olist_products table on the product_id column.
+-- 4. Filter the results to only include orders where the order status is 'delivered'.
+-- 5. Group the results by the order ID.
+-- 6. Order the results by the order ID.
+SELECT
+    ooi.order_id,
+    SUM(ooi.freight_value) AS freight_value,
+    SUM(op.product_weight_g) AS product_weight_g
+FROM
+    olist_orders o
+    JOIN olist_order_items ooi ON o.order_id = ooi.order_id
+    JOIN olist_products op ON ooi.product_id = op.product_id
+WHERE
+    o.order_status = 'delivered'
+GROUP BY
+    ooi.order_id
+ORDER BY
+    ooi.order_id

sql/global_amount_order_status.sql

@@ -0,0 +1,13 @@
+-- Calculates the amount of orders for each order status.
+--
+-- Explanation step by step:
+-- 1. Select the order status and the amount of orders for each order status.
+-- 2. Join the olist_orders table with the olist_order_items table on the order_id column.
+-- 3. Group the results by the order status.
+SELECT
+    oo.order_status,
+    COUNT(oo.order_status) AS Ammount
+FROM
+    olist_orders oo
+GROUP BY
+    oo.order_status;

sql/orders_per_day_and_holidays_2017.sql

@@ -0,0 +1,26 @@
+-- Calculates the number of orders per day and whether each day is a holiday.
+--
+-- Explanation step by step:
+-- 1. Select the number of orders per day and whether each day is a holiday.
+-- 2. Join the olist_orders table with the public_holidays table on the date column.
+-- 3. Filter the results to only include orders from 2017.
+-- 4. Group the results by the date.
+-- 5. Order the results by the date.
+SELECT
+    COUNT(o.order_id) AS order_count,
+    CAST(
+        STRFTIME ('%s', DATE(o.order_purchase_timestamp)) AS INTEGER
+    ) * 1000 AS date,
+    CASE
+        WHEN DATE(h.date) IS NOT NULL THEN 'true'
+        ELSE 'false'
+    END AS holiday
+FROM
+    olist_orders o
+    LEFT JOIN public_holidays h ON DATE(o.order_purchase_timestamp) = DATE(h.date)
+WHERE
+    STRFTIME ('%Y', o.order_purchase_timestamp) = '2017'
+GROUP BY
+    DATE(o.order_purchase_timestamp)
+ORDER BY
+    DATE(o.order_purchase_timestamp);

sql/real_vs_estimated_delivered_time.sql

@@ -0,0 +1,93 @@
+-- Calculate the real and estimated delivery time for each month
+--
+-- It will have different columns: 
+-- 1. month_no, with the month numbers going from 01 to 12
+-- 2. month, with the 3 first letters of each month (e.g. Jan, Feb)
+-- 3. Year2016_real_time, with the average delivery time per month of 2016 (NaN if it doesn't exist) 
+-- 4. Year2017_real_time, with the average delivery time per month of 2017 (NaN if it doesn't exist)
+-- 5. Year2018_real_time, with the average delivery time per month of 2018 (NaN if it doesn't exist)
+-- 6. Year2016_estimated_time, with the average estimated delivery time per month of 2016 (NaN if it doesn't exist) 
+-- 7. Year2017_estimated_time, with the average estimated delivery time per month of 2017 (NaN if it doesn't exist)
+-- 8. Year2018_estimated_time, with the average estimated delivery time per month of 2018 (NaN if it doesn't exist).
+--
+-- Explanation step by step:
+-- 1. Calculate the real and estimated delivery time for each order
+-- 2. Group the data by month
+-- 3. Calculate the average real and estimated delivery time for each month
+WITH
+    base AS (
+        SELECT
+            STRFTIME ('%m', oo.order_purchase_timestamp) AS month_no,
+            STRFTIME ('%Y', oo.order_purchase_timestamp) AS year,
+            julianday (oo.order_delivered_customer_date) - julianday (oo.order_purchase_timestamp) AS real_time,
+            julianday (oo.order_estimated_delivery_date) - julianday (oo.order_purchase_timestamp) AS estimated_time
+        FROM
+            olist_orders oo
+        WHERE
+            oo.order_status = 'delivered'
+            AND oo.order_delivered_customer_date IS NOT NULL
+    ),
+    pivot AS (
+        SELECT
+            b.month_no,
+            AVG(
+                CASE
+                    WHEN year = '2016' THEN b.real_time
+                END
+            ) AS Year2016_real_time,
+            AVG(
+                CASE
+                    WHEN year = '2017' THEN b.real_time
+                END
+            ) AS Year2017_real_time,
+            AVG(
+                CASE
+                    WHEN year = '2018' THEN b.real_time
+                END
+            ) AS Year2018_real_time,
+            AVG(
+                CASE
+                    WHEN year = '2016' THEN b.estimated_time
+                END
+            ) AS Year2016_estimated_time,
+            AVG(
+                CASE
+                    WHEN year = '2017' THEN b.estimated_time
+                END
+            ) AS Year2017_estimated_time,
+            AVG(
+                CASE
+                    WHEN year = '2018' THEN b.estimated_time
+                END
+            ) AS Year2018_estimated_time
+        FROM
+            base b
+        GROUP BY
+            month_no
+    )
+SELECT
+    p.month_no,
+    CASE p.month_no
+        WHEN '01' THEN 'Jan'
+        WHEN '02' THEN 'Feb'
+        WHEN '03' THEN 'Mar'
+        WHEN '04' THEN 'Apr'
+        WHEN '05' THEN 'May'
+        WHEN '06' THEN 'Jun'
+        WHEN '07' THEN 'Jul'
+        WHEN '08' THEN 'Aug'
+        WHEN '09' THEN 'Sep'
+        WHEN '10' THEN 'Oct'
+        WHEN '11' THEN 'Nov'
+        WHEN '12' THEN 'Dec'
+    END AS month,
+    p.Year2016_real_time,
+    p.Year2017_real_time,
+    p.Year2018_real_time,
+    p.Year2016_estimated_time,
+    p.Year2017_estimated_time,
+    p.Year2018_estimated_time
+FROM
+    pivot p
+ORDER BY
+    p.month_no;

sql/revenue_by_month_year.sql

@@ -0,0 +1,126 @@
+-- Calculates revenue by month and year
+--
+-- It will have different columns: 
+-- 1. month_no, with the month numbers going from 01 to 12
+-- 2. month, with the 3 first letters of each month (e.g. Jan, Feb)
+-- 3. Year2016, with the revenue per month of 2016 (0.00 if it doesn't exist)
+-- 4. Year2017, with the revenue per month of 2017 (0.00 if it doesn't exist)
+-- 5. Year2018, with the revenue per month of 2018 (0.00 if it doesn't exist)
+--
+-- Explanation step by step:
+-- 1. Calculate the revenue for each order
+-- 2. Group the data by month
+-- 3. Calculate the average revenue for each month
+WITH
+    month_names AS (
+        SELECT
+            '01' AS month_no,
+            'Jan' AS month
+        UNION ALL
+        SELECT
+            '02',
+            'Feb'
+        UNION ALL
+        SELECT
+            '03',
+            'Mar'
+        UNION ALL
+        SELECT
+            '04',
+            'Apr'
+        UNION ALL
+        SELECT
+            '05',
+            'May'
+        UNION ALL
+        SELECT
+            '06',
+            'Jun'
+        UNION ALL
+        SELECT
+            '07',
+            'Jul'
+        UNION ALL
+        SELECT
+            '08',
+            'Aug'
+        UNION ALL
+        SELECT
+            '09',
+            'Sep'
+        UNION ALL
+        SELECT
+            '10',
+            'Oct'
+        UNION ALL
+        SELECT
+            '11',
+            'Nov'
+        UNION ALL
+        SELECT
+            '12',
+            'Dec'
+    ),
+    -- Get the minimum payment per order
+    min_payments AS (
+        SELECT
+            oop.order_id,
+            MIN(oop.payment_value) AS min_payment
+        FROM
+            olist_order_payments oop
+        GROUP BY
+            oop.order_id
+    ),
+    -- Calculate revenue grouped by year and month
+    revenue AS (
+        SELECT
+            strftime ('%m', oo.order_delivered_customer_date) AS month_no,
+            strftime ('%Y', oo.order_delivered_customer_date) AS year,
+            SUM(mp.min_payment) AS total_revenue
+        FROM
+            olist_orders oo
+            JOIN min_payments mp ON oo.order_id = mp.order_id
+        WHERE
+            oo.order_status = 'delivered'
+            AND oo.order_delivered_customer_date IS NOT NULL
+            AND strftime ('%Y', oo.order_delivered_customer_date) IN ('2016', '2017', '2018')
+        GROUP BY
+            month_no,
+            year
+    )
+    -- Final Select
+SELECT
+    mn.month_no,
+    mn.month,
+    COALESCE(
+        MAX(
+            CASE
+                WHEN r.year = '2016' THEN r.total_revenue
+            END
+        ),
+        0.0
+    ) AS Year2016,
+    COALESCE(
+        MAX(
+            CASE
+                WHEN r.year = '2017' THEN r.total_revenue
+            END
+        ),
+        0.0
+    ) AS Year2017,
+    COALESCE(
+        MAX(
+            CASE
+                WHEN r.year = '2018' THEN r.total_revenue
+            END
+        ),
+        0.0
+    ) AS Year2018
+FROM
+    month_names mn
+    LEFT JOIN revenue r ON mn.month_no = r.month_no
+GROUP BY
+    mn.month_no,
+    mn.month
+ORDER BY
+    mn.month_no;

sql/revenue_per_state.sql

@@ -0,0 +1,28 @@
+-- Calculate the revenue per state
+--
+-- It will have different columns: 
+-- 1. customer_state, with the state of the customer
+-- 2. Revenue, with the revenue per state
+--
+-- Explanation step by step:
+-- 1. Calculate the revenue for each order
+-- 2. Group the data by state
+-- 3. Calculate the average revenue for each state
+-- 4. Order the data by revenue
+-- 5. Limit the data to the top 10
+SELECT
+    oc.customer_state AS customer_state,
+    SUM(oop.payment_value) AS Revenue
+FROM
+    olist_orders oo
+    JOIN olist_customers oc ON oo.customer_id = oc.customer_id
+    JOIN olist_order_payments oop ON oop.order_id = oo.order_id
+WHERE
+    oo.order_status = 'delivered'
+    AND oo.order_delivered_customer_date IS NOT NULL
+GROUP BY
+    oc.customer_state
+ORDER BY
+    Revenue DESC
+LIMIT
+    10;

sql/top_10_least_revenue_categories.sql

@@ -0,0 +1,33 @@
+-- Calculates the top 10 least revenue categories
+--
+-- It will have different columns: 
+-- 1. Category, with the category name
+-- 2. Num_order, with the number of orders
+-- 3. Revenue, with the revenue
+--
+-- Explanation step by step:
+-- 1. Calculate the revenue for each order
+-- 2. Group the data by category
+-- 3. Calculate the average revenue for each category
+-- 4. Order the data by revenue
+-- 5. Limit the data to the top 10
+SELECT
+    pcnt.product_category_name_english AS Category,
+    COUNT(DISTINCT oo.order_id) AS Num_order,
+    SUM(p.payment_value) AS Revenue
+FROM
+    olist_orders oo
+    JOIN olist_order_items ooi ON oo.order_id = ooi.order_id
+    JOIN olist_products op ON ooi.product_id = op.product_id
+    JOIN product_category_name_translation pcnt ON op.product_category_name = pcnt.product_category_name
+    JOIN olist_order_payments p ON oo.order_id = p.order_id
+WHERE
+    oo.order_status = 'delivered'
+    AND oo.order_delivered_customer_date IS NOT NULL
+    AND op.product_category_name IS NOT NULL
+GROUP BY
+    Category
+ORDER BY
+    Revenue ASC
+LIMIT
+    10;

sql/top_10_revenue_categories.sql

@@ -0,0 +1,33 @@
+-- Calculates the top 10 revenue categories
+--  
+-- It will have different columns:
+-- 1. Category, with the category name
+-- 2. Num_order, with the number of orders  
+-- 3. Revenue, with the revenue
+--
+-- Explanation step by step:
+-- 1. Calculate the revenue for each order
+-- 2. Group the data by category
+-- 3. Calculate the average revenue for each category
+-- 4. Order the data by revenue
+-- 5. Limit the data to the top 10
+SELECT
+    pcnt.product_category_name_english AS Category,
+    COUNT(DISTINCT oo.order_id) AS Num_order,
+    SUM(p.payment_value) AS Revenue
+FROM
+    olist_orders oo
+    JOIN olist_order_items ooi ON oo.order_id = ooi.order_id
+    JOIN olist_products op ON ooi.product_id = op.product_id
+    JOIN product_category_name_translation pcnt ON op.product_category_name = pcnt.product_category_name
+    JOIN olist_order_payments p ON oo.order_id = p.order_id
+WHERE
+    oo.order_status = 'delivered'
+    AND oo.order_delivered_customer_date IS NOT NULL
+    AND op.product_category_name IS NOT NULL
+GROUP BY
+    Category
+ORDER BY
+    Revenue DESC
+LIMIT
+    10;