Level 2 · The bar that looks four times taller

From Daniel Foster, Mobile Product Manager

“Marketing's deck has this chart and the caption says desktop converts 'several times better' than mobile, so they want to move budget away from the app. The bars do look dramatic. Is the gap really that big?”

Your goal

The chart's y axis starts at 2% (0.02), not at zero. Set result to a dict with "looks": how many times taller the desktop bar is drawn than the mobile bar, (desktop - 0.02) / (mobile - 0.02), and "real": the real ratio of the two conversion rates, desktop / mobile. Use the conversion rate (share of sessions with converted True) per device from web_sessions, and round both to 2 decimals.

Use the web_sessions DataFramenot done yet
Store a dict in resultnot done yet

Conversion rate by device (axis starts at 2%)

The chart from the deck. Look at where the y axis begins.

Show the numbers

	conversion_rate
desktop	3.9%
tablet	2.8%
mobile	2.4%

Your code

Python

Loading the editor…

Reference solution

ResultPreview of web_sessions · press Run to see yours

Preview of web_sessions
session_id	customer_id	started_at	device	source
1	NaN	2023-01-01 23:02:24	mobile	paid
2	18	2023-01-01 15:07:49	desktop	paid
3	NaN	2023-01-01 16:55:46	desktop	direct
4	NaN	2023-01-01 12:43:28	mobile	organic

Mote checks your result

Press Run. Each row that matches the goal becomes a firefly in your lantern.

What Daniel gets, and why

The same rates, axis from zero

Drawn from zero, each bar's length is proportional to its rate: desktop is about 1.6 times mobile.

Show the numbers

	conversion_rate
desktop	3.9%
tablet	2.8%
mobile	2.4%

Next: Six lines or six small charts

DataFrames

Loaded for you as pandas DataFrames, plus pd and np.

web_sessions40,000 rows · 7 columns

session_idinteger
customer_idinteger
started_attimestamp
devicevarchar
sourcevarchar
pages_viewedinteger
convertedboolean

First 5 rows of web_sessions
session_id	customer_id	started_at	device	source	pages_viewed	converted
1	NaN	2023-01-01 23:02:24	mobile	paid	1	false
2	18	2023-01-01 15:07:49	desktop	paid	1	false
3	NaN	2023-01-01 16:55:46	desktop	direct	7	false
4	NaN	2023-01-01 12:43:28	mobile	organic	3	false
5	NaN	2023-01-01 13:06:21	mobile	social	2	false

customers2,018 rows · 9 columns

customer_idinteger
first_namevarchar
last_namevarchar
emailvarchar
cityvarchar
statevarchar
signup_datedate
segmentvarchar
referral_sourcevarchar

First 5 rows of customers
customer_id	first_name	last_name	email	city	state	signup_date	segment	referral_source
1	Owen	Hudson	owen.hudson@icloud.com	San Francisco	CA	2022-06-02	consumer	search
2	Douglas	Mendoza	dmendoza69@gmail.com	Atlanta	GA	2022-06-02	consumer	NaN
3	Henry	Brooks	hbrooks@lakesidecafe.com	Los Angeles	CA	2022-06-04	business	NaN
4	Brian	Graham	bgraham@bluecedarhomes.com	Seattle	WA	2022-06-06	business	pinterest
5	Kyle	Jimenez	kjimenez27@yahoo.com	Indianapolis	IN	2022-06-07	consumer	NaN

products120 rows · 8 columns

product_idinteger
namevarchar
categoryvarchar
subcategoryvarchar
unit_pricedecimal
unit_costdecimal
launched_ondate
is_activeboolean

First 5 rows of products
product_id	name	category	subcategory	unit_price	unit_cost	launched_on	is_active
101	Vale Velvet Loveseat	Furniture	Sofas	1659.00	NaN	2023-09-11	true
102	Alder Oak Coffee Table	Furniture	Tables	919.00	503.50	2024-05-16	true
103	Orchard Rattan Dining Chair	Furniture	Chairs	789.00	429.87	2023-05-27	true
104	Upland Oak Bookcase	Furniture	Storage	429.00	214.70	2022-12-20	true
105	Avery Leather Bed Frame	Furniture	Beds	449.00	232.70	2021-08-31	true

orders12,000 rows · 7 columns

order_idinteger
customer_idinteger
order_datetimestamp
statusvarchar
channelvarchar
shipping_feedecimal
coupon_codevarchar

First 5 rows of orders
order_id	customer_id	order_date	status	channel	shipping_fee	coupon_code
1001	125	2023-01-01 20:51:19	completed	web	0.00	NaN
1002	156	2023-01-02 07:58:46	completed	web	0.00	NaN
1003	22	2023-01-02 10:25:27	completed	marketplace	4.99	WELCOME10
1004	48	2023-01-02 11:21:51	completed	marketplace	4.99	WELCOME10
1005	82	2023-01-02 11:43:36	completed	web	0.00	WELCOME10

order_items30,338 rows · 6 columns

order_item_idinteger
order_idinteger
product_idinteger
quantityinteger
unit_pricedecimal
discount_pctdecimal

First 5 rows of order_items
order_item_id	order_id	product_id	quantity	unit_price
1	1001	121	9	31.99
2	1001	154	3	11.99
3	1002	154	4	11.99
4	1002	194	1	149.00
5	1002	193	2	19.99

returns1,673 rows · 5 columns

return_idinteger
order_item_idinteger
returned_ondate
reasonvarchar
refund_amountdecimal

First 5 rows of returns
return_id	order_item_id	returned_on	reason	refund_amount
1	14	2023-01-10	damaged	83.69
2	42	2023-01-10	wrong_item	19.99
3	105	2023-01-14	wrong_item	10.99
4	24	2023-01-16	changed_mind	26.99
5	126	2023-01-16	late	20.99

employees60 rows · 8 columns

employee_idinteger
first_namevarchar
last_namevarchar
departmentvarchar
titlevarchar
manager_idinteger
hire_datedate
salaryinteger

First 5 rows of employees
employee_id	first_name	last_name	department	title	manager_id	hire_date	salary
1	Catherine	Hale	Executive	Chief Executive Officer	NaN	2016-03-01	285000
2	Robert	Kim	Executive	Chief Operating Officer	1	2017-01-09	235000
3	Megan	Collins	Sales	Head of Sales	1	2018-04-16	172000
4	Chloe	Gutierrez	Sales	Account Executive	3	2020-01-21	72000
5	Frank	Bell	Sales	Account Executive	3	2023-05-21	73000

marketing_spend3,276 rows · 3 columns

spend_datedate
channelvarchar
amountdecimal

First 5 rows of marketing_spend
spend_date	channel	amount
2023-01-01	paid	492.99
2023-01-01	email	37.38
2023-01-01	social	200.53
2023-01-02	paid	709.13
2023-01-02	email	64.13

The lesson

Bars are read by length

When people look at a bar chart they compare how long the bars are. That only works if every bar starts from the same place, and that place is zero. Then a bar twice as long means a value twice as large, which is exactly what the eye assumes.

Cut the axis off at 2% and every bar loses its first 2%. What’s left is not the value but the value minus 2%, and the ratio between two of those leftovers can be anything.

Measuring the distortion

The conversion rate is the share of sessions that ended in an order. converted is True or False, and the mean of a True/False column is the share of Trues:

rates = web_sessions.groupby("device")["converted"].mean()
# desktop 0.0387, tablet 0.0281, mobile 0.0241

The real ratio is desktop / mobile. The drawn ratio is (desktop - 0.02) / (mobile - 0.02), because only the part above the axis start is drawn. The closer the smaller value sits to the axis start, the more extreme the drawn ratio gets: move the start up to 2.2% and the same data looks almost eight times taller.

Is the gap real?

Yes. Desktop converts better than mobile on every traffic source in this data, and 40,000 sessions make a 1.5-point gap far bigger than chance. The chart’s problem is not the direction, it’s the size. “Desktop converts 1.6 times as often as mobile” is true. “Several times better” is what the axis made up. A budget decision built on the second sentence is built on the axis.

When a non-zero axis is fine

Line charts are read by slope and position, not length, so a line chart of daily revenue doesn’t need to start at zero, and forcing it there can flatten real changes. The rule is about bars and areas: anything whose length or filled area is the message must start at zero.

Common mistakes

Trusting the default axis. Some tools pick the axis range from the data. Check where a bar chart’s axis starts before you paste it into a deck.
Fixing the chart by adding a note. Readers see the bars before they read the footnote. Redraw it from zero.
Describing ratios as differences, or the other way round. “1.5 points higher” and “60% higher” are both true here. Say which one you mean.
Comparing rates without the sizes behind them. A rate from 300 sessions is much noisier than one from 20,000. Chapter 3 covers how sure a rate is.