Records dashboard

Demo. Every figure below is invented.

Show me a Signed in as

Owner view. You can see the money: the cost, margin and fee figures below, and the two right-hand columns of every table. Supervisor view. The cost, margin and fee figures are gone — one summary tile and two table columns. They are not hidden on screen; the database never sends them to this account, so there is nothing to find in the page.

A sheeting and bag-making plant. Reels in, sheets and bags out. The board follows the material and what it costs to lose it.

Jobs this month

148

Paper in

212.4

tonnes

Sellable out

190.9

tonnes

Yield

89.9%

target 92%▲ Below target

Margin this month

19.4%

across all jobs

Yield by month

Sellable output as a share of paper in

This yearLast year

Every month this year has come in under target, and the gap widened from June.

Yield by month (%)
Month	This year	Last year
Jan	91.8	90.2
Feb	92.1	90.8
Mar	91.4	91.1
Apr	90.9	90.6
May	91.2	91.5
Jun	90.4	91.0
Jul	89.8	90.7
Aug	90.1	90.9
Sep	89.2	91.2
Oct	88.7	90.4
Nov	89.9	90.8
Dec	89.4	91.1

Waste by machine

Share of input lost as trim and spoilage

▲ above target

Sheeter 2 loses nearly twice what Sheeter 1 does on the same grades.

Waste by machine (%)
Item	Value
Sheeter 1	6.1
Sheeter 2	11.4
Printer	7.6
Bag line A	5.2
Bag line B	6.8

Downtime by cause

Hours stopped, this month

Reel change alone accounts for 32% of the total.

Downtime by cause
Cause	Value
Reel change	42
Web break	31
Setup	24
Glue jam	18
Power cut	9
Other	6

Recent jobs

Every row keeps the scan of the original page beside it.

The last two columns reach the owner only

Job	Date	Product	In (kg)	Out (kg)	Yield	Cost/t	Margin	Source
J-4821	08 Sep	Kraft bag 12 kg	1,840	1,663	90.4%	12,450	18%	View scan
J-4820	08 Sep	Sheets 70×100 90 g	2,310	2,142	92.7%	11,900	22%	View scan
J-4819	07 Sep	Kraft bag 5 kg	1,120	982	87.7%	13,100	11%	View scan
J-4818	07 Sep	Sheets 64×90 80 g	3,050	2,769	90.8%	11,750	21%	View scan
J-4817	06 Sep	Printed bag, 2 colour	980	846	86.3%	14,220	9%	View scan
J-4816	06 Sep	Sheets 70×100 120 g	2,640	2,441	92.5%	11,880	23%	View scan

Crescendo Academy — head of school viewCrescendo Academy — supervisor view

A school teaching Python to its seconde classes. These are the screens the teacher and the head see, in the platform's own colours. The board follows one class through one module, step by step, and asks the students themselves which step was hard.

Students on the platform

108

five classes

Average progress

62%

across all classes

Stuck more than 3 days

students, all classes▲ Needs follow-up

Exam questions changed

this term, 2 after students answered▲ Needs follow-up

Fees outstanding

412,000

EGP, this term

Every class, not only this one

What the head of studies opens first: the same module, five classes, and where each of them actually is

The fee column reaches the owner only

Seconde C is fifteen points behind Seconde B on the same module with the same material, and four of its students have not moved in more than three days. That is a conversation with one teacher, not a curriculum problem — and it is the kind of thing that stays invisible until every class is on one screen.

Every class, not only this one
Class	Students	Progress	Stuck
Seconde A	24	64%	3
Seconde B	26	71%	2
Seconde C	22	49%	4
Première — spé maths	19	58%	1
Terminale — NSI	17	66%	1

Where each student is

Seconde A, in detail · Mathématiques — Module 1 : variables et calculs · one cell per step

doneon it nowstuck > 3 daysnot reached

This is the screen that replaces asking. A teacher can see in one look who has finished, who is working, and who has been on the same step for more than three days — which is the only one of the three that needs them today.

Where each student is
Student	Steps done	Progress	Days on it
Farida Hosny	20	100%	—
Nour Adel	20	100%	—
Mariam Saad	20	100%	—
Yara Fahmy	19	95%	2
Habiba Selim	18	90%	1
Karim Wassef	17	85%	2
Sandra Nabil	16	80%	1
Clara Ghali	16	80%	2
Omar Fahmy	15	75%	2
Dina Rached	14	70%	9
Salma Aziz	14	70%	2
Nadine Boutros	14	70%	—
Maya Chedid	14	70%	1
Laila Mansour	14	70%	—
Youssef Nabil	13	65%	2
Rita Sedky	12	60%	—
Malak Fouad	11	55%	5
Aya Gamal	10	50%	—
Camille Rizk	9	45%	1
Lina Adly	7	35%	1
Perrine Khalil	6	30%	—
Hana Sobhy	5	25%	7
Verena Saad	3	15%	2
Rana Tewfik	0	0%	—

Where the class stops

Students who reached each step of the module

the biggest drop

The cliff is step 14, “Demander une valeur” — 5 students got that far and no further. That is a step to teach again before the class moves on, and it is invisible in any class average.

Where the class stops
Step	Students
1. Qu'est-ce qu'une variable ?	23
2. Écrire sa première variable	23
3. Afficher avec print()	23
4. À toi : afficher un message	22
5. Les nombres entiers	22
6. Les nombres décimaux	21
7. À toi : calculer une moyenne	20
8. Addition et soustraction	19
9. Multiplication et division	19
10. À toi : le prix total	18
11. La division entière	17
12. Le reste d'une division	16
13. À toi : pair ou impair ?	15
14. Demander une valeur	14
15. Convertir un texte en nombre	9
16. À toi : demander l'âge	8
17. Arrondir un résultat	6
18. Afficher une phrase complète	5
19. À toi : la facture	4
20. Bilan du module	3

What the students said about each step

One optional click at the end of a step, worst received first. It never blocks the lesson, so about half the class answers — and only steps with four answers or more are shown here.

confusedhardcleareasy

169 answers in all, on a question nobody has to answer. 2 steps cross the line where more than half of the students who answered found them confusing or hard, worst of all step 16, “À toi : demander l'âge”. Step 14 is in that list too — it is also the step the class stops at in the board above, which is two different signals pointing at the same lesson.

What the students said about each step
Step	Confused	Hard	Clear	Easy	Answers
16. À toi : demander l'âge	3	4	0	1	8
14. Demander une valeur	2	4	2	0	8
7. À toi : calculer une moyenne	3	2	2	3	10
10. À toi : le prix total	0	5	3	2	10
13. À toi : pair ou impair ?	3	2	4	1	10
18. Afficher une phrase complète	0	2	2	0	4
4. À toi : afficher un message	1	3	4	2	10
15. Convertir un texte en nombre	1	1	1	2	5

Who changed what, and whether anyone had already answered

Every change to a question, a mark scheme or a student's attempts — kept with its author and its time, and impossible to remove

changed after students had submitted

14 Sep, 16:42 Mme Chedid reworded question 3 Contrôle 2 · Module 1 after 18 of 24 had submitted
13 Sep, 09:10 Mme Chedid granted 2 extra attempts to Youssef Nabil Exercice 14 · Demander une valeur before any submission
12 Sep, 11:20 M. Rizk reopened the exam for Hana Sobhy until 16 Sep Contrôle 1 · Module 1 before any submission
11 Sep, 18:05 Mme Chedid changed question 5 from 4 points to 3 Contrôle 2 · Module 1 before any submission
10 Sep, 08:55 M. Rizk granted the whole class 1 extra attempt Exercice 10 · À toi : le prix total before any submission
08 Sep, 15:30 Mme Chedid deleted question 7 Contrôle 1 · Module 1 after marking had started

Six changes this term and two of them landed after students had answered — which is exactly the case a paper exam cannot answer at all. Nothing here can be edited or removed, by anyone, including the person who made the change and including the head of the school.

Seconde A, this module

Every row opens that student's own history — every step, every attempt.

Fees due and last payment reach the owner only

Student	Class	Step	Progress	Days on it	Last active	Fees due	Last paid	Student page
Farida Hosny	Seconde A	finished	100%	—	today	0	02 Sep	Open student
Mariam Saad	Seconde A	finished	100%	—	today	1,200	21 Aug	Open student
Omar Fahmy	Seconde A	15 / 20	75%	2	2 days ago	0	28 Aug	Open student
Maya Chedid	Seconde A	14 / 20	70%	1	yesterday	4,500	12 Jul	Open student
Hana Sobhy	Seconde A	5 / 20	25%	7	7 days ago	600	15 Aug	Open student
Rana Tewfik	Seconde A	0 / 20	0%	—	never	4,500	—	Open student

A firm filing on behalf of clients. The board follows the invoices both ways, the deadlines, and who is owed leave.

Active clients

212

Invoices out, unpaid

worth 214,500 EGP◆ Watch

Filed on time

96.5%

target 98%◆ Watch

Leave days left

across the team

Fee income

186,400

EGP, this month

Invoices in and invoices out

Issued against received, EGP thousands

Issued to clientsReceived from suppliers▲ marked month: more came in than went out

Twice this year the bills coming in were larger than the invoices going out. Both times it was a month with a big corporate-tax job still unbilled.

Invoices in and invoices out (k)
Month	Issued to clients	Received from suppliers
Jan	182	131
Feb	164	172
Mar	191	138
Apr	205	144
May	198	150
Jun	212	147
Jul	178	186
Aug	194	141
Sep	186	139
Oct	201	148
Nov	190	152
Dec	196	150

Leave taken and days left

Against each person's annual entitlement

taken▲ 15+ days left

Two people have barely taken leave this year. At full entitlement that is a liability on the books and a December with nobody in the office.

Leave taken and days left (days)
Person	Entitled	Taken	Left
Rania Fouad	21	21	0
Sameh Ragab	21	19	2
Amir Sobhy	21	11	10
Tarek Aziz	25	9	16
Hoda Mansour	21	4	17
Dina Kamel	21	2	19

Why a filing was rejected

Count, this year

Missing signature alone accounts for 37% of the total.

Why a filing was rejected
Cause	Value
Missing signature	34
Wrong period	21
Mismatched total	17
Late submission	12
Other	7

Turnaround by service

Working days from documents received to filed

▲ above target

Corporate tax takes three times longer than VAT and drives the late filings.

Turnaround by service (d)
Item	Value
VAT return	2.1
Payroll	1.4
Corporate tax	6.8
Audit prep	4.2
Registration	3

Invoices in and out

Every row keeps the scan of the original invoice beside it.

Amount and margin reach the owner only

Invoice	Date	Party	Direction	Status	Amount (EGP)	Margin	Source
INV-1042	08 Sep	Client C-0412	Out	Sent	3,800	41%	View scan
BIL-0391	07 Sep	Cairo Print House	In	Paid	1,240	—	View scan
INV-1041	05 Sep	Client C-0411	Out	Paid	24,000	28%	View scan
INV-1040	04 Sep	Client C-0409	Out	Overdue 14 d	3,800	12%	View scan
BIL-0390	02 Sep	Office rent, September	In	Paid	9,500	—	View scan
INV-1039	01 Sep	Client C-0408	Out	Sent	18,500	31%	View scan

A lab with years of run sheets, instrument logs and sample tracking on paper. The board carries the results themselves, not just the throughput.

Samples logged

3,480

this year

Genes over threshold

148

q < 0.05, latest comparison

Runs passing QC

91.4%

target 95%▲ Below target

Samples reaching a figure

56%

of everything logged◆ Watch

Reagent cost / run

1,240

EGP, this month

Expression by gene and condition

log2 fold change against the untreated control, latest comparison

downup

The first four genes move together and with dose; MKI67 moves the other way, which is the proliferation arrest. One cell does not fit its row — HSPA1A at 48 h — and that is either real or a plate mix-up. The scan of the run sheet is one click away in the table below, which is the whole point of keeping them.

Expression by gene and condition (log2 fold change)
Gene	6 h	12 h	24 h	48 h	Dose ×2	Dose ×4
IL6	+1.2	+2.4	+2.9	+2.1	+3.1	+3.4
TNF	+0.8	+1.9	+2.2	+1.4	+2.6	+2.8
CXCL8	+1.5	+2.1	+2.6	+1.9	+2.9	+3.0
NFKB1	+0.4	+0.9	+1.1	+0.6	+1.3	+1.5
SOD2	+0.2	+0.5	+0.9	+1.2	+1.1	+1.6
HSPA1A	-0.1	+0.3	+1.4	+2.2	+1.8	+2.4
CAT	-0.3	-0.8	-1.2	-1.6	-1.4	-2.1
MKI67	-0.6	-1.4	-2.2	-2.8	-2.5	-3.2

Association across the genome

Every variant tested, plotted as −log10 of its p-value

clears significancep = 5 × 10⁻⁸

Four loci clear the line, and the strongest sits in the HLA region on chromosome 6 — where immune studies tend to land. Everything under the line is noise until it replicates, which is exactly why the run sheet and the sample provenance behind each point have to stay reachable. That is the same argument as the scans beside the rows below.

Loci clearing genome-wide significance
Nearest gene	Chr	Mb	−log10 p	p
HLA-DRB1	6	32.6	9.6	3e-10
CDKN2B-AS1	9	22.1	8.4	4e-09
FTO	16	53.8	7.9	1e-08
IL1B	2	112.8	7.6	3e-08

From sample to figure

Where this year's 3,480 samples ended up

right column: lost at that step

Nearly a third of the samples that clear every technical step never reach a figure. That is not an instrument problem, and while the sheets sat in folders nobody could see it at all.

From sample to figure
Stage	Samples	Share
Samples logged	3,480	100%
Passed intake QC	3,180	91%
Sequenced	3,012	87%
Aligned, Q30 and up	2,884	83%
Analysed	2,750	79%
Used in a figure	1,940	56%

Failed runs by instrument

Share of runs needing a repeat

▲ above target

Sequencer B fails more than twice as often as Sequencer A on the same assays.

Failed runs by instrument (%)
Item	Value
Sequencer A	4.2
Sequencer B	9.8
qPCR 1	3.1
qPCR 2	5.4
Plate reader	2.6

Why a run was discarded

Count, this year

Low yield alone accounts for 37% of the total.

Why a run was discarded
Cause	Value
Low yield	38
Contamination	26
Instrument error	19
Sample degraded	14
Operator error	7

Recent runs

Every row keeps the scan of the original page beside it.

The last two columns reach the owner only

Run	Date	Assay	Samples	Passed	QC	Cost/run	Reagent lot	Source
R-2214	08 Sep	16S amplicon	96	91	94.8%	1,180	LOT-8841	View scan
R-2213	08 Sep	qPCR panel	48	47	97.9%	640	LOT-8839	View scan
R-2212	07 Sep	WGS, low input	24	20	83.3%	3,420	LOT-8836	View scan
R-2211	07 Sep	16S amplicon	96	88	91.7%	1,180	LOT-8841	View scan
R-2210	06 Sep	qPCR panel	48	46	95.8%	640	LOT-8839	View scan
R-2209	06 Sep	WGS, low input	24	22	91.7%	3,420	LOT-8836	View scan

A charity running five appeals at once, with money arriving by bank transfer, a card page and a cash box. The board follows how close each appeal is and where the donors are.

Donors this year

2,146

2,880 gifts

Average gift

787

EGP

Campaign targets met

70%

average across five◆ Watch

Repeat donors

38%

gave more than once

Cost to raise 100 EGP

14.20

EGP, this year

How close each appeal is

Raised against target, with the time left

raised▲ past targetbar length is the target, so the campaigns are comparable

School fees is the one to worry about: 486,000 EGP short with 41 days to run, which is 11,900 a day from here. The clinic appeal went past its target and that surplus is sitting unallocated.

How close each appeal is (EGP)
Campaign	Target	Raised	Share	Days left
Ramadan food boxes	1,200,000	1,104,000	92%	23
School fees, 120 children	900,000	414,000	46%	41
Clean water, 3 villages	600,000	168,000	28%	58
Winter blankets	400,000	318,000	80%	12
Clinic equipment	250,000	262,000	105%	6

Where the donors are

Donors this year, by country

More than half the donors are in the United States and Canada — 1,210 of 2,146. The work happens in Egypt and the money comes from North America, which is an argument about the card page, the currency and the language of the receipt, not about the cause.

Where the donors are
Country	Donors	Share
United States	742	34.6%
Canada	468	21.8%
Egypt	396	18.5%
France	214	10.0%
UAE	128	6.0%
Germany	88	4.1%
Other	110	5.1%
Total	2,146	100%

What a gift is actually worth

Gifts this year, by size, EGP

The average gift is 787 EGP and almost nobody gives 787 EGP. The 64 donors in the last band gave a quarter of everything raised — they are the list to know by name.

What a gift is actually worth (count)
Gift size	Gifts
Under 200	964
200–499	810
500–999	604
1,000–4,999	438
5,000+	64

Average gift by month

What the typical donation is worth

This yearLast year

Ramadan and the December appeal carry the year. In the eight months between them the average gift sits below the 800 the budget assumes.

Average gift by month ( EGP)
Month	This year	Last year
Jan	768	742
Feb	1240	1180
Mar	1096	1020
Apr	792	798
May	795	776
Jun	786	784
Jul	774	762
Aug	758	748
Sep	787	770
Oct	762	752
Nov	858	840
Dec	934	906

Recent donations

Every row keeps the scan of the bank slip or the receipt beside it.

Amount and cost to raise reach the owner only

Receipt	Date	Donor	Appeal	Country	Channel	Amount (EGP)	Cost to raise	Source
R-20418	09 Sep	Hanna Sedky	Ramadan food boxes	Egypt	Bank transfer	5,000	1.2%	View scan
R-20417	09 Sep	Anonymous	School fees	United States	Card online	1,800	3.4%	View scan
R-20416	08 Sep	M. Ghali	Clinic equipment	France	Bank transfer	12,000	0.8%	View scan
R-20415	08 Sep	Youssef Bakr	Winter blankets	Egypt	Cash, office	300	—	View scan
R-20414	07 Sep	Diaspora circle, 12 gifts	Clean water	Canada	Card online	9,400	3.6%	View scan
R-20413	07 Sep	N. Abdelmalek	School fees	UAE	Bank transfer	2,500	1.1%	View scan

Five organisations, one system underneath — and five different boards, because a laboratory does not need a leave calendar and a charity does not need a waste chart. Switch the role to watch the money disappear: that is enforced in the database, not in the page. How this gets built →

Month	Issued to clients	Received from suppliers
Jan	182	131
Feb	164	172
Mar	191	138
Apr	205	144
May	198	150
Jun	212	147
Jul	178	186
Aug	194	141
Sep	186	139
Oct	201	148
Nov	190	152
Dec	196	150

Month	Issued to clients	Received from suppliers
Jan	182	131
Feb	164	172
Mar	191	138
Apr	205	144
May	198	150
Jun	212	147
Jul	178	186
Aug	194	141
Sep	186	139
Oct	201	148
Nov	190	152
Dec	196	150

Month	Issued to clients	Received from suppliers
Jan	182	131
Feb	164	172
Mar	191	138
Apr	205	144
May	198	150
Jun	212	147
Jul	178	186
Aug	194	141
Sep	186	139
Oct	201	148
Nov	190	152
Dec	196	150