Create stephlinj answers.sql

sql-assessment/answers.sql

@@ -0,0 +1,148 @@
+-- Q1
+SELECT
+    date,
+    SUM(impressions) AS total_impressions
+FROM
+    `pmg-sql-assessment.datasets.marketing_performance`
+GROUP BY
+    date
+ORDER BY
+    date;
+
+-- Q2
+WITH StateRevenue AS (
+    SELECT
+        state,
+        SUM(revenue) AS total_revenue
+    FROM
+        `pmg-sql-assessment.datasets.website_revenue`
+    GROUP BY
+        state
+)
+
+SELECT
+    state,
+    total_revenue
+FROM
+    StateRevenue
+ORDER BY
+    total_revenue DESC
+LIMIT 3
+OFFSET 2;
+
+-- The third best state (OH) generated $37577 in revenue.
+
+-- Q3
+WITH CampaignSummary AS (
+    SELECT
+        ci.name AS campaign_name,
+        mp.campaign_id,
+        ROUND(SUM(mp.cost), 2) AS total_cost,
+        SUM(mp.impressions) AS total_impressions,
+        SUM(mp.clicks) AS total_clicks,
+        SUM(wr.revenue) AS total_revenue
+    FROM
+        `pmg-sql-assessment.datasets.campaign_info` AS ci
+    LEFT JOIN
+        `pmg-sql-assessment.datasets.marketing_performance` AS mp
+    ON
+        ci.id = mp.campaign_id
+    LEFT JOIN
+        `pmg-sql-assessment.datasets.website_revenue` AS wr
+    ON
+        ci.id = wr.campaign_id
+    GROUP BY
+        campaign_name, campaign_id
+)
+
+SELECT
+    campaign_name,
+    total_cost,
+    total_impressions,
+    total_clicks,
+    total_revenue
+FROM
+    CampaignSummary;
+
+
+-- Q4
+SELECT
+    mp.geo AS state,
+    SUM(mp.conversions) AS total_conversions
+FROM
+    `pmg-sql-assessment.datasets.marketing_performance` AS mp
+JOIN
+    `pmg-sql-assessment.datasets.campaign_info` AS ci
+ON
+    mp.campaign_id = ci.id
+WHERE
+    ci.name = 'Campaign5'
+GROUP BY
+    state
+ORDER BY
+    total_conversions DESC;
+
+-- Georgia generated the most conversions with a total of 672.
+
+-- Q5
+WITH CampaignMetrics AS (
+    SELECT
+        ci.name AS campaign_name,
+        SUM(mp.conversions) AS total_conversions,
+        SUM(mp.cost) AS total_cost
+    FROM
+        `pmg-sql-assessment.datasets.campaign_info` AS ci
+    LEFT JOIN
+        `pmg-sql-assessment.datasets.marketing_performance` AS mp
+    ON
+        ci.id = mp.campaign_id
+    GROUP BY
+        campaign_name
+)
+
+SELECT
+    campaign_name,
+    SUM(total_conversions) AS total_conversions,
+    SUM(total_cost) AS total_cost,
+    SUM(total_conversions) / SUM(total_cost) AS roas
+FROM
+    CampaignMetrics
+GROUP BY
+    campaign_name
+ORDER BY
+    roas DESC;
+
+-- Looking at efficiency of marketing campaigns, we can look at return on ad spend (ROAS). This calculates the revenue generated per unit of advertising. Looking at the results, Campaign 4 has the highest ROAS of about 2.34 compared to the other four campaigns.
+
+-- Bonus
+WITH DayOfWeekMetrics AS (
+    SELECT
+        EXTRACT(DAYOFWEEK FROM date) AS day_of_week,
+        SUM(clicks) AS total_clicks,
+        SUM(impressions) AS total_impressions
+    FROM
+        `pmg-sql-assessment.datasets.marketing_performance`
+    GROUP BY
+        day_of_week
+)
+
+SELECT
+    CASE
+        WHEN day_of_week = 1 THEN 'Sunday'
+        WHEN day_of_week = 2 THEN 'Monday'
+        WHEN day_of_week = 3 THEN 'Tuesday'
+        WHEN day_of_week = 4 THEN 'Wednesday'
+        WHEN day_of_week = 5 THEN 'Thursday'
+        WHEN day_of_week = 6 THEN 'Friday'
+        WHEN day_of_week = 7 THEN 'Saturday'
+        ELSE 'Unknown'
+    END AS day_of_week,
+    ROUND(AVG(total_clicks / total_impressions), 3) AS avg_ctr
+FROM
+    DayOfWeekMetrics
+GROUP BY
+    day_of_week
+ORDER BY
+    avg_ctr DESC;
+
+-- The best day of the week to run ads would be Wednesday with the highest average click through rate of about 1.098.