Spaces:
Paused
Paused
Update eda.py
Browse files
eda.py
CHANGED
|
@@ -95,7 +95,7 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 95 |
FROM orders
|
| 96 |
GROUP BY order_status
|
| 97 |
ORDER BY count DESC
|
| 98 |
-
")
|
| 99 |
orders_monthly = sql_df(conn, """
|
| 100 |
SELECT date(strftime('%Y-%m-01', order_purchase_timestamp)) AS month,
|
| 101 |
COUNT(*) AS count
|
|
@@ -103,7 +103,7 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 103 |
WHERE order_purchase_timestamp IS NOT NULL
|
| 104 |
GROUP BY date(strftime('%Y-%m-01', order_purchase_timestamp))
|
| 105 |
ORDER BY month
|
| 106 |
-
")
|
| 107 |
delivery = sql_df(conn, """
|
| 108 |
SELECT AVG(
|
| 109 |
julianday(order_delivered_customer_date) - julianday(order_purchase_timestamp)
|
|
@@ -111,7 +111,7 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 111 |
FROM orders
|
| 112 |
WHERE order_delivered_customer_date IS NOT NULL
|
| 113 |
AND order_purchase_timestamp IS NOT NULL
|
| 114 |
-
")
|
| 115 |
delivery_hist_raw = sql_df(conn, """
|
| 116 |
SELECT
|
| 117 |
CAST(julianday(order_delivered_customer_date) - julianday(order_purchase_timestamp) AS INTEGER) AS days
|
|
@@ -119,7 +119,7 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 119 |
WHERE order_delivered_customer_date IS NOT NULL
|
| 120 |
AND order_purchase_timestamp IS NOT NULL
|
| 121 |
AND (julianday(order_delivered_customer_date) - julianday(order_purchase_timestamp)) BETWEEN -1 AND 60
|
| 122 |
-
")
|
| 123 |
if not delivery_hist_raw.empty:
|
| 124 |
bins = pd.IntervalIndex.from_breaks(range(-1, 62, 1))
|
| 125 |
s = pd.cut(delivery_hist_raw["days"], bins=bins)
|
|
@@ -140,7 +140,7 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 140 |
FROM order_payments
|
| 141 |
GROUP BY payment_type
|
| 142 |
ORDER BY count DESC
|
| 143 |
-
")
|
| 144 |
else:
|
| 145 |
payments = pd.DataFrame(columns=["payment_type","count"])
|
| 146 |
|
|
@@ -150,14 +150,14 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 150 |
FROM order_reviews
|
| 151 |
GROUP BY review_score
|
| 152 |
ORDER BY review_score
|
| 153 |
-
")
|
| 154 |
review_text_stats = sql_df(conn, """
|
| 155 |
SELECT
|
| 156 |
AVG(LENGTH(COALESCE(review_comment_message,'')) - LENGTH(REPLACE(COALESCE(review_comment_message,''),' ','')) + 1) AS avg_len_words,
|
| 157 |
SUM(CASE WHEN LENGTH(COALESCE(review_comment_message,'')) > 0 THEN 1 ELSE 0 END) AS n_texts,
|
| 158 |
AVG(review_score) AS avg_review_score
|
| 159 |
FROM order_reviews
|
| 160 |
-
")
|
| 161 |
texts = sql_df(conn, "SELECT review_comment_message FROM order_reviews WHERE review_comment_message IS NOT NULL LIMIT 5000")
|
| 162 |
if not texts.empty:
|
| 163 |
terms = (
|
|
@@ -190,18 +190,18 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 190 |
|
| 191 |
def categories_join():
|
| 192 |
if has_products and has_trans:
|
| 193 |
-
return """
|
| 194 |
SELECT p.product_id,
|
| 195 |
COALESCE(t.product_category_name_english, p.product_category_name) AS cat_en
|
| 196 |
FROM products p
|
| 197 |
LEFT JOIN product_category_name_translation t
|
| 198 |
ON t.product_category_name = p.product_category_name
|
| 199 |
-
"
|
| 200 |
elif has_products:
|
| 201 |
-
return "
|
| 202 |
SELECT product_id, product_category_name AS cat_en
|
| 203 |
FROM products
|
| 204 |
-
"
|
| 205 |
else:
|
| 206 |
return None
|
| 207 |
|
|
@@ -245,7 +245,7 @@ def build_summaries_from_db(db_path: str) -> dict:
|
|
| 245 |
WHERE o.order_delivered_customer_date IS NOT NULL AND o.order_purchase_timestamp IS NOT NULL
|
| 246 |
GROUP BY c.customer_state
|
| 247 |
ORDER BY avg_delay DESC
|
| 248 |
-
")
|
| 249 |
else:
|
| 250 |
delays_by_state = pd.DataFrame(columns=["customer_state","avg_delay"])
|
| 251 |
|
|
@@ -549,4 +549,4 @@ def render_eda():
|
|
| 549 |
if dataset_choice in ["All", "Products"]:
|
| 550 |
show_products()
|
| 551 |
if dataset_choice in ["All", "Customers"]:
|
| 552 |
-
show_customers()
|
|
|
|
| 95 |
FROM orders
|
| 96 |
GROUP BY order_status
|
| 97 |
ORDER BY count DESC
|
| 98 |
+
""")
|
| 99 |
orders_monthly = sql_df(conn, """
|
| 100 |
SELECT date(strftime('%Y-%m-01', order_purchase_timestamp)) AS month,
|
| 101 |
COUNT(*) AS count
|
|
|
|
| 103 |
WHERE order_purchase_timestamp IS NOT NULL
|
| 104 |
GROUP BY date(strftime('%Y-%m-01', order_purchase_timestamp))
|
| 105 |
ORDER BY month
|
| 106 |
+
""")
|
| 107 |
delivery = sql_df(conn, """
|
| 108 |
SELECT AVG(
|
| 109 |
julianday(order_delivered_customer_date) - julianday(order_purchase_timestamp)
|
|
|
|
| 111 |
FROM orders
|
| 112 |
WHERE order_delivered_customer_date IS NOT NULL
|
| 113 |
AND order_purchase_timestamp IS NOT NULL
|
| 114 |
+
""")
|
| 115 |
delivery_hist_raw = sql_df(conn, """
|
| 116 |
SELECT
|
| 117 |
CAST(julianday(order_delivered_customer_date) - julianday(order_purchase_timestamp) AS INTEGER) AS days
|
|
|
|
| 119 |
WHERE order_delivered_customer_date IS NOT NULL
|
| 120 |
AND order_purchase_timestamp IS NOT NULL
|
| 121 |
AND (julianday(order_delivered_customer_date) - julianday(order_purchase_timestamp)) BETWEEN -1 AND 60
|
| 122 |
+
""")
|
| 123 |
if not delivery_hist_raw.empty:
|
| 124 |
bins = pd.IntervalIndex.from_breaks(range(-1, 62, 1))
|
| 125 |
s = pd.cut(delivery_hist_raw["days"], bins=bins)
|
|
|
|
| 140 |
FROM order_payments
|
| 141 |
GROUP BY payment_type
|
| 142 |
ORDER BY count DESC
|
| 143 |
+
""")
|
| 144 |
else:
|
| 145 |
payments = pd.DataFrame(columns=["payment_type","count"])
|
| 146 |
|
|
|
|
| 150 |
FROM order_reviews
|
| 151 |
GROUP BY review_score
|
| 152 |
ORDER BY review_score
|
| 153 |
+
""")
|
| 154 |
review_text_stats = sql_df(conn, """
|
| 155 |
SELECT
|
| 156 |
AVG(LENGTH(COALESCE(review_comment_message,'')) - LENGTH(REPLACE(COALESCE(review_comment_message,''),' ','')) + 1) AS avg_len_words,
|
| 157 |
SUM(CASE WHEN LENGTH(COALESCE(review_comment_message,'')) > 0 THEN 1 ELSE 0 END) AS n_texts,
|
| 158 |
AVG(review_score) AS avg_review_score
|
| 159 |
FROM order_reviews
|
| 160 |
+
""")
|
| 161 |
texts = sql_df(conn, "SELECT review_comment_message FROM order_reviews WHERE review_comment_message IS NOT NULL LIMIT 5000")
|
| 162 |
if not texts.empty:
|
| 163 |
terms = (
|
|
|
|
| 190 |
|
| 191 |
def categories_join():
|
| 192 |
if has_products and has_trans:
|
| 193 |
+
return """""
|
| 194 |
SELECT p.product_id,
|
| 195 |
COALESCE(t.product_category_name_english, p.product_category_name) AS cat_en
|
| 196 |
FROM products p
|
| 197 |
LEFT JOIN product_category_name_translation t
|
| 198 |
ON t.product_category_name = p.product_category_name
|
| 199 |
+
"""
|
| 200 |
elif has_products:
|
| 201 |
+
return """
|
| 202 |
SELECT product_id, product_category_name AS cat_en
|
| 203 |
FROM products
|
| 204 |
+
"""
|
| 205 |
else:
|
| 206 |
return None
|
| 207 |
|
|
|
|
| 245 |
WHERE o.order_delivered_customer_date IS NOT NULL AND o.order_purchase_timestamp IS NOT NULL
|
| 246 |
GROUP BY c.customer_state
|
| 247 |
ORDER BY avg_delay DESC
|
| 248 |
+
""")
|
| 249 |
else:
|
| 250 |
delays_by_state = pd.DataFrame(columns=["customer_state","avg_delay"])
|
| 251 |
|
|
|
|
| 549 |
if dataset_choice in ["All", "Products"]:
|
| 550 |
show_products()
|
| 551 |
if dataset_choice in ["All", "Customers"]:
|
| 552 |
+
show_customers()
|