Seller Analysis

Let’s create a helper function.

def seller_top(metric: str, show_cnt: bool=True, ascending=False):
    """Show Top Customers by Metric"""
    cols = ['seller_id', metric]
    if show_cnt:
        if metric == 'products_cnt':
            cols += ['orders_cnt']
        else:
            cols += ['products_cnt', 'orders_cnt']
    display(
        df_sellers[cols]
        .sort_values(metric, ascending=ascending)
        .set_index('seller_id')
        .head(10)
    )

Number of Products

Let’s identify the top sellers.

seller_top('products_cnt')

	products_cnt	orders_cnt
seller_id
6560211a19b47992c3666cc44a7e94c0	2,003.00	1,826.00
4a3ca9315b744ce9f8e9374361493884	1,952.00	1,775.00
1f50f920176fa81dab994f9023523100	1,926.00	1,399.00
cc419e0650a3c5ba77189a1882b7556a	1,719.00	1,651.00
da8622b14eb17ae2831f4ac5b9dab84a	1,548.00	1,311.00
955fee9216a65b617aa5c0531780ce60	1,475.00	1,264.00
1025f0e2d44d7041d6cf58b6550e0bfa	1,420.00	910.00
7c67e1448b00f6e969d365cea6b010ab	1,356.00	974.00
ea8482cd71df3c1969d7b9473ff13abc	1,188.00	1,132.00
7a67c85e85bb2ce8582c35f2203ad736	1,156.00	1,146.00

Key Observations:

• Seller “6560211a19b47992c3666cc44a7e94c0” sold the most products

Let’s see at statistics and distribution of the metric.

df_sellers['products_cnt'].explore.info(
    labels=dict(unique_products_cnt='Number of Products')
    , title='Distribution of Number of Products Per Seller'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "products_cnt" (Type: Integer)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	2.00k		Mean	37.33		1	471 (15%)
Missing	148 (5%)		99%	437.26		Trimmed Mean (10%)	15.06		2	278 (9%)
Distinct	262 (8%)		95%	151		Mode	1		3	193 (6%)
Non-Duplicate	114 (4%)		75%	26		Range	2.00k		4	145 (5%)
Duplicates	2.83k (92%)		50%	8		IQR	24		5	143 (5%)
Dup. Values	148 (5%)		25%	2		Std	120.08		6	106 (3%)
Zeros	---		5%	1		MAD	10.38		7	97 (3%)
Negative	---		1%	1		Kurt	119.73		8	95 (3%)
Memory Usage	<1 Mb		Min	1		Skew	9.61		9	79 (3%)

Key Observations:

• 75% of sellers sold ≤26 products total

• Top 5% sold >150 products

Number of Unique Products

Let’s identify the top sellers.

seller_top('unique_products_cnt')

	unique_products_cnt	products_cnt	orders_cnt
seller_id
4a3ca9315b744ce9f8e9374361493884	395.00	1,952.00	1,775.00
d91fb3b7d041e83b64a00a3edfb37e4f	309.00	527.00	526.00
cca3071e3e9bb7d12640c9fbe2301306	307.00	804.00	691.00
fa1c13f2614d7b5c4749cbc52fecda94	285.00	581.00	580.00
7142540dd4c91e2237acb7e911c4eba2	265.00	328.00	316.00
6560211a19b47992c3666cc44a7e94c0	254.00	2,003.00	1,826.00
da8622b14eb17ae2831f4ac5b9dab84a	221.00	1,548.00	1,311.00
ea8482cd71df3c1969d7b9473ff13abc	219.00	1,188.00	1,132.00
3d871de0142ce09b7081e2b9d1733cb1	202.00	1,131.00	1,064.00
7c67e1448b00f6e969d365cea6b010ab	197.00	1,356.00	974.00

Key Observations:

• Seller “4a3ca9315b744ce9f8e9374361493884” sold the most unique products

Let’s see at statistics and distribution of the metric.

df_sellers['unique_products_cnt'].explore.info(
    labels=dict(unique_products_cnt='Number of Unique Products')
    , title='Distribution of Number of Unique Products Per Seller'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "unique_products_cnt" (Type: Integer)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	395		Mean	11.39		1	669 (22%)
Missing	148 (5%)		99%	126		Trimmed Mean (10%)	6.14		2	414 (13%)
Distinct	127 (4%)		95%	46		Mode	1		3	285 (9%)
Non-Duplicate	49 (2%)		75%	10		Range	394		4	207 (7%)
Duplicates	2.97k (96%)		50%	4		IQR	8		5	166 (5%)
Dup. Values	78 (3%)		25%	2		Std	24.55		6	123 (4%)
Zeros	---		5%	1		MAD	4.45		8	102 (3%)
Negative	---		1%	1		Kurt	60.57		7	101 (3%)
Memory Usage	<1 Mb		Min	1		Skew	6.49		9	88 (3%)

Key Observations:

• 75% of sellers sold ≤10 unique products

• Top 5% sold >45 unique products

Number of Orders

Let’s identify the top sellers.

seller_top('orders_cnt', show_cnt=False)

	orders_cnt
seller_id
6560211a19b47992c3666cc44a7e94c0	1,826.00
4a3ca9315b744ce9f8e9374361493884	1,775.00
cc419e0650a3c5ba77189a1882b7556a	1,651.00
1f50f920176fa81dab994f9023523100	1,399.00
da8622b14eb17ae2831f4ac5b9dab84a	1,311.00
955fee9216a65b617aa5c0531780ce60	1,264.00
7a67c85e85bb2ce8582c35f2203ad736	1,146.00
ea8482cd71df3c1969d7b9473ff13abc	1,132.00
4869f7a5dfa277a7dca6462dcf3b52b2	1,124.00
3d871de0142ce09b7081e2b9d1733cb1	1,064.00

Key Observations:

• Seller “6560211a19b47992c3666cc44a7e94c0” participated in the most orders

Let’s see at statistics and distribution of the metric.

df_sellers['orders_cnt'].explore.info(
    labels=dict(orders_cnt='Number of Orders')
    , title='Distribution of Number of Orders Per Seller'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "orders_cnt" (Type: Integer)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	1.83k		Mean	33.15		1	517 (17%)
Missing	148 (5%)		99%	402.78		Trimmed Mean (10%)	13.27		2	307 (10%)
Distinct	247 (8%)		95%	133		Mode	1		3	209 (7%)
Non-Duplicate	112 (4%)		75%	22		Range	1.82k		4	153 (5%)
Duplicates	2.85k (92%)		50%	7		IQR	20		5	141 (5%)
Dup. Values	135 (4%)		25%	2		Std	105.79		6	110 (4%)
Zeros	---		5%	1		MAD	8.90		7	96 (3%)
Negative	---		1%	1		Kurt	117.00		8	94 (3%)
Memory Usage	<1 Mb		Min	1		Skew	9.43		9	84 (3%)

Key Observations:

• 75% of sellers participated in ≤22 orders

• Top 5% participated in ≥130 orders

Total Sales Revenue

Let’s identify the top sellers.

seller_top('revenue')

	revenue	products_cnt	orders_cnt
seller_id
4869f7a5dfa277a7dca6462dcf3b52b2	226,987.93	1,148.00	1,124.00
53243585a1d6dc2643021fd1853d8905	218,620.44	404.00	352.00
4a3ca9315b744ce9f8e9374361493884	197,210.82	1,952.00	1,775.00
fa1c13f2614d7b5c4749cbc52fecda94	191,785.04	581.00	580.00
7c67e1448b00f6e969d365cea6b010ab	186,740.04	1,356.00	974.00
7e93a43ef30c4f03f38b393420bc753a	164,582.49	321.00	318.00
da8622b14eb17ae2831f4ac5b9dab84a	159,816.87	1,548.00	1,311.00
7a67c85e85bb2ce8582c35f2203ad736	139,908.68	1,156.00	1,146.00
1025f0e2d44d7041d6cf58b6550e0bfa	138,208.56	1,420.00	910.00
955fee9216a65b617aa5c0531780ce60	132,138.71	1,475.00	1,264.00

Key Observations:

• Seller “4869f7a5dfa277a7dca6462dcf3b52b2” generated the most revenue

Let’s see at statistics and distribution of the metric.

df_sellers['revenue'].explore.info(
    labels=dict(revenue='Seller Revenue')
    , title='Distribution of Seller Revenue'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "revenue" (Type: Float)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	226.99k		Mean	4.48k		59.90	7 (<1%)
Missing	148 (5%)		99%	54.44k		Trimmed Mean (10%)	1.84k		69.90	7 (<1%)
Distinct	2.68k (86%)		95%	16.97k		Mode	Multiple		120	7 (<1%)
Non-Duplicate	2.52k (81%)		75%	3.51k		Range	226.98k		119.90	6 (<1%)
Duplicates	418 (14%)		50%	851.10		IQR	3.28k		180	6 (<1%)
Dup. Values	157 (5%)		25%	228.70		Std	13.96k		49.90	6 (<1%)
Zeros	---		5%	49		MAD	1.11k		29.90	6 (<1%)
Negative	---		1%	17.77		Kurt	102.35		89.90	6 (<1%)
Memory Usage	<1 Mb		Min	6.50		Skew	8.87		79.90	6 (<1%)

Key Observations:

• 75% of sellers made ≤3,500 R$

• Top 5% made ≥17,000 R$

Number of Products per Order

Let’s identify the top sellers.

seller_top('avg_prouducts_cnt')

	avg_prouducts_cnt	products_cnt	orders_cnt
seller_id
0b36063d5818f81ccb94b54adfaebbf5	15.00	15.00	1.00
99a25c39b28a74d1151c35c18d178292	6.00	6.00	1.00
a3b42d266fa8afc874b909422ce88582	6.00	6.00	1.00
c394e193cda3b4225ff2094d32184849	5.50	11.00	2.00
010da0602d7774602cd1b3f5fb7b709e	5.00	5.00	1.00
8ff38bc3969e67c36c48343a07090f66	4.00	8.00	2.00
fc6295add6f51a0936407ead70c1001d	4.00	4.00	1.00
ec8463980a4e0ea9f8517aea1ed0c419	4.00	8.00	2.00
b37c4c02bda3161a7546a4e6d222d5b2	3.75	15.00	4.00
ab27bbbad5239bc31a34709275a70db4	3.67	11.00	3.00

Key Observations:

• Seller ‘0b36063d5818f81ccb94b54adfaebbf5’ has highest average products per order (single order)

Let’s see at statistics and distribution of the metric.

df_sellers['avg_prouducts_cnt'].explore.info(
    labels=dict(avg_prouducts_cnt='Average Number of Products in Order')
    , title='Distribution of Average Number of Products in Order per Sellers'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "avg_prouducts_cnt" (Type: Float)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	15		Mean	1.16		1	1.56k (51%)
Missing	148 (5%)		99%	3		Trimmed Mean (10%)	1.07		1.50	69 (2%)
Distinct	511 (17%)		95%	1.82		Mode	1		2	62 (2%)
Non-Duplicate	359 (12%)		75%	1.14		Range	14		1.33	57 (2%)
Duplicates	2.58k (83%)		50%	1		IQR	0.14		1.17	42 (1%)
Dup. Values	152 (5%)		25%	1		Std	0.44		1.25	41 (1%)
Zeros	---		5%	1		MAD	0		1.14	38 (1%)
Negative	---		1%	1		Kurt	346.25		1.11	28 (<1%)
Memory Usage	<1 Mb		Min	1		Skew	13.39		1.20	28 (<1%)

Key Observations:

• 75% of sellers average 1.14 products per order

• Top 1% average ≥3 products

Total Product Value per Order

Let’s identify the top sellers.

seller_top('avg_order_total_price')

	avg_order_total_price	products_cnt	orders_cnt
seller_id
e3b4998c7a498169dc7bce44e6bb6277	6,735.00	1.00	1.00
80ceebb4ee9b31afb6c6a916a574a1e2	6,729.00	1.00	1.00
ee27a8f15b1dded4d213a468ba4eb391	6,499.00	1.00	1.00
b37c4c02bda3161a7546a4e6d222d5b2	6,018.75	15.00	4.00
585175ec331ea177fa47199e39a6170a	3,549.00	1.00	1.00
abe021b01ba992245271b9aa422032df	3,360.00	2.00	2.00
a00824eb9093d40e589b940ec45c4eb0	3,133.32	3.00	3.00
e2a1ac9bf33e5549a2a4f834e70df2f8	2,999.89	5.00	5.00
e908c0f3646e8b60375734a350d95d71	2,951.00	1.00	1.00
d63c73efd41eb002280e7ec831424edb	2,799.00	2.00	2.00

Key Observations:

• Sellers “e3b4998c7a498169dc7bce44e6bb6277” and “80ceebb4ee9b31afb6c6a916a574a1e2” had highest order values (single order each)

Let’s see at statistics and distribution of the metric.

df_sellers['avg_order_total_price'].explore.info(
    labels=dict(avg_order_total_price='Average Amount of Products in Order, R$')
    , title='Distribution of Average Amount of Products in Order Per Seller'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "avg_order_total_price" (Type: Float)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	6.74k		Mean	196.49		79.90	11 (<1%)
Missing	148 (5%)		99%	1.61k		Trimmed Mean (10%)	127.87		59.90	11 (<1%)
Distinct	2.63k (85%)		95%	641.15		Mode	Multiple		69.90	11 (<1%)
Non-Duplicate	2.47k (80%)		75%	189		Range	6.73k		29.90	9 (<1%)
Duplicates	466 (15%)		50%	105.53		IQR	128.72		39.90	9 (<1%)
Dup. Values	157 (5%)		25%	60.28		Std	369.92		49.90	8 (<1%)
Zeros	---		5%	26.17		MAD	80.72		119.90	7 (<1%)
Negative	---		1%	14.23		Kurt	127.40		109.90	6 (<1%)
Memory Usage	<1 Mb		Min	6		Skew	9.16		90	6 (<1%)

Key Observations:

• 75% of sellers average ≤189 R$ per order

• Top 5% average ≥641 R$

Product Price per Order

Let’s identify the top sellers.

seller_top('avg_product_price')

	avg_product_price	products_cnt	orders_cnt
seller_id
e3b4998c7a498169dc7bce44e6bb6277	6,735.00	1.00	1.00
80ceebb4ee9b31afb6c6a916a574a1e2	6,729.00	1.00	1.00
ee27a8f15b1dded4d213a468ba4eb391	6,499.00	1.00	1.00
585175ec331ea177fa47199e39a6170a	3,549.00	1.00	1.00
abe021b01ba992245271b9aa422032df	3,360.00	2.00	2.00
a00824eb9093d40e589b940ec45c4eb0	3,133.32	3.00	3.00
e2a1ac9bf33e5549a2a4f834e70df2f8	2,999.89	5.00	5.00
e908c0f3646e8b60375734a350d95d71	2,951.00	1.00	1.00
d63c73efd41eb002280e7ec831424edb	2,799.00	2.00	2.00
1444c08e64d55fb3c25f0f09c07ffcf2	2,749.00	1.00	1.00

Key Observations:

• Seller “e3b4998c7a498169dc7bce44e6bb6277” has highest average product price (single order)

Let’s see at statistics and distribution of the metric.

df_sellers['avg_product_price'].explore.info(
    labels=dict(avg_product_price='Average Product Price in Order, R$')
    , title='Distribution of Average Product Price in Order Per Seller'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "avg_product_price" (Type: Float)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	6.74k		Mean	180.46		59.90	14 (<1%)
Missing	148 (5%)		99%	1.58k		Trimmed Mean (10%)	116.39		69.90	12 (<1%)
Distinct	2.58k (83%)		95%	595.09		Mode	59.90		79.90	11 (<1%)
Non-Duplicate	2.42k (78%)		75%	174		Range	6.73k		29.90	11 (<1%)
Duplicates	516 (17%)		50%	97.24		IQR	119.77		99	10 (<1%)
Dup. Values	163 (5%)		25%	54.23		Std	347.44		39.90	10 (<1%)
Zeros	---		5%	23.90		MAD	76.36		119.90	8 (<1%)
Negative	---		1%	12.90		Kurt	138.90		120	8 (<1%)
Memory Usage	<1 Mb		Min	6		Skew	9.45		49.90	8 (<1%)

Key Observations:

• 75% of sellers average ≤174 R$ per product

• Top 5% average ≥595 R$

Product Weight

Let’s identify the top sellers.

seller_top('avg_product_weight_kg')

	avg_product_weight_kg	products_cnt	orders_cnt
seller_id
5d043cd5512d4bd2f88e5ccdd5736c38	30.00	1.00	1.00
d7827b2af99326a03b0ed9c7a24db0d3	30.00	1.00	1.00
731ef20c231d9a7103a425e83fd91271	30.00	3.00	2.00
19a7c9f66ffa6452ff4b4c0cca71da27	30.00	1.00	1.00
a99504ac23ed125aaf5302af6cc30af9	30.00	1.00	1.00
a9415cf907dfa12f9ed18693244f5329	30.00	2.00	1.00
08cdbae123ff67ca4e36d9d641ce0119	30.00	6.00	4.00
ccb83a794700270fde70898fe9ff368b	30.00	12.00	12.00
6fa9202c10491e472dffd59a3e82b2a3	30.00	3.00	3.00
a70e9066dbfd10b6b97a3f54a1356762	30.00	4.00	2.00

Key Observations:

• Maximum average product weight: 30kg

Let’s see at statistics and distribution of the metric.

df_sellers['avg_product_weight_kg'].explore.info(
    labels=dict(avg_product_weight_kg='Average Weight of Products, kg')
    , title='Distribution of Average Weight of Products Per Seller'
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "avg_product_weight_kg" (Type: Float)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	30		Mean	2.68		0.20	69 (2%)
Missing	148 (5%)		99%	24.05		Trimmed Mean (10%)	1.65		0.30	57 (2%)
Distinct	767 (25%)		95%	11.14		Mode	0.20		0.50	46 (1%)
Non-Duplicate	383 (12%)		75%	2.73		Range	29.95		0.40	46 (1%)
Duplicates	2.33k (75%)		50%	1		IQR	2.27		0.15	46 (1%)
Dup. Values	384 (12%)		25%	0.46		Std	4.35		0.60	35 (1%)
Zeros	---		5%	0.19		MAD	1.04		0.35	34 (1%)
Negative	---		1%	0.10		Kurt	14.02		0.55	33 (1%)
Memory Usage	<1 Mb		Min	0.05		Skew	3.40		0.25	31 (1%)

Key Observations:

• 75% of sellers average ≤2.7kg

• Top 5% average ≥11kg

Carrier Handoff Delay

Let’s identify the top sellers.

seller_top('avg_carrier_delivery_delay_days')

	avg_carrier_delivery_delay_days	products_cnt	orders_cnt
seller_id
586a871d4f1221763fddb6ceefdeb95e	45.43	2.00	1.00
8e670472e453ba34a379331513d6aab1	40.15	1.00	1.00
20f0aeea30bc3b8c4420be8ced4226c0	32.60	1.00	1.00
e09887ca8c7bf8a4621ce481820414ef	32.02	3.00	2.00
6d04126aba80df143fd038e711b8fd96	26.87	1.00	1.00
244b04680fdbded0acf5aebd9c92b44a	23.73	2.00	2.00
8d92f3ea807b89465643c219455e7369	21.52	8.00	3.00
8629a7efec1aab257e58cda559f03ba7	19.31	1.00	1.00
bd4ff04377f974574f7f0bf9d6ce7dde	19.10	1.00	1.00
c004e5ea15737026cecaee0447e00b75	18.52	1.00	1.00

Key Observations:

• Seller “586a871d4f1221763fddb6ceefdeb95e” had maximum carrier handoff delay: 45 days

Let’s see at statistics and distribution of the metric.

df_sellers['avg_carrier_delivery_delay_days'].explore.info(
    labels=dict(avg_carrier_delivery_delay_days='Average Carrier Delivery Delay Time, days')
    , title='Distribution of Average Carrier Delivery Delay Time Per Seller'
    , lower_quantile=0.05
    , upper_quantile=0.95
    , hist_mode='dual_hist_trim'
)

Summary Statistics for "avg_carrier_delivery_delay_days" (Type: Float)

Summary			Percentiles			Detailed Stats			Value Counts
Total	2.95k (95%)		Max	45.43		Mean	-3.04		-0.39	1 (<1%)
Missing	149 (5%)		99%	9.34		Trimmed Mean (10%)	-3.24		-5.22	1 (<1%)
Distinct	2.95k (95%)		95%	1.17		Mode	Multiple		-3.00	1 (<1%)
Non-Duplicate	2.95k (95%)		75%	-2.20		Range	79.55		-3.52	1 (<1%)
Duplicates	148 (5%)		50%	-3.28		IQR	2.06		-4.50	1 (<1%)
Dup. Values	0 (<1%)		25%	-4.26		Std	3.55		-9.99	1 (<1%)
Zeros	---		5%	-6.51		MAD	1.53		-2.93	1 (<1%)
Negative	2.72k (88%)		1%	-11.68		Kurt	37.56		-3.35	1 (<1%)
Memory Usage	<1 Mb		Min	-34.12		Skew	3.03		-4.03	1 (<1%)

Key Observations:

• Top 5% of sellers delivered to carrier ≥6.5 days early

• 75% delivered ≥2 days early

• Bottom 5% delayed ≥1 day