Online coding and AI classes in the Tumbes region

INEI, Censos Nacionales 2025, tabulated demographic indicators, Cuadro N° 1, people counted plus those INEI estimates were missed, reference day 15 September 2025. The department figure comes from INEI's own table; we do not total the provinces.

School numbers are rising, unlike in several Andean regions. INEI's compendium shows public secundaria enrolment in Tumbes climbing from 14.7 thousand in 2015 to 21.4 thousand in 2025, and public primaria from 23.0 thousand to 26.4 thousand. The department counted 4,953 teachers in 2025.

INEI household survey (ENAHO) results printed in the Compendio Estadístico Perú 2026, first volume, tables 11.5, 11.6 and 11.13. Decimals shown with points.

If there is one laptop in the house, it is enough. Younger children can do block coding on a tablet; from roughly ten we ask for a laptop or desktop, shared if necessary. The free class checks what you have before anyone pays.

import csv

with open("gastos-gore-tumbes.csv", newline="", encoding="utf-8") as f:
    amounts = [float(row["IMPORTE"]) for row in csv.DictReader(f)]

width = max(amounts) / 10
bins = [0] * 10
for x in amounts:
    bins[min(int(x // width), 9)] += 1

print("Rows:", len(amounts))
for i, n in enumerate(bins):
    print("bin", i, n)

Version 1 divides the range from zero to the largest amount into ten equal slices. The result is useless as a picture: 19,088 of 19,624 rows, about 97 per cent, land in the first slice, and seven of the ten bars are empty or nearly so. One or two very large values stretched the scale until everything else was squashed into a single bar. When numbers vary by factors of ten, bin them by order of magnitude instead: one-digit amounts together, two-digit amounts together, and so on.

import csv
import math

with open("gastos-gore-tumbes.csv", newline="", encoding="utf-8") as f:
    amounts = [float(row["IMPORTE"]) for row in csv.DictReader(f)]

by_log = {}
by_digits = {}
for x in amounts:
    a = int(math.log(x, 10))
    b = len(str(int(x))) - 1
    by_log[a] = by_log.get(a, 0) + 1
    by_digits[b] = by_digits.get(b, 0) + 1

for power in sorted(by_digits):
    bar = "#" * round(by_digits[power] / 500)
    print(power, by_log.get(power, 0), by_digits[power], bar)
disagree = [x for x in amounts if int(math.log(x, 10)) != len(str(int(x))) - 1]
print("Rows where the two methods disagree:", len(disagree))
print("math.log(1000, 10) =", math.log(1000, 10))

Now the shape appears: most payments have four digits, with smaller groups either side. But look at the second and third columns. The logarithm method puts 3,996 rows in the three-digit bin and the digit-counting method 3,902. The 94 rows that differ are all amounts of exactly one thousand. math.log(1000, 10) is computed by dividing two natural logarithms, and the tiny rounding error leaves it just under 3, so int() drops it to 2. Counting digits, or using math.log10, which returns exactly 3.0 for 1000, files them correctly. The bars use the digit counts.

Data: Gobierno Regional de Tumbes, Gastos según fuentes financiamiento del Gobierno Regional de Tumbes, ODC-By. Counts and shares are ours; shares are rounded to one decimal place, so they need not add to exactly 100. We print no individual amount and say nothing about any payment's purpose.

1. Under 8, with coins and notes. Sort play money into piles by how many digits it shows. Which pile is tallest?
2. Blocks, around 8 to 10. Make a list of twenty numbers between 3 and 30,000. Sort them into bins by digit count, and draw the bars.
3. First Python, around 10 to 12. Print len(str(n)) for 7, 70, 700 and 7000. Predict each answer first.
4. Ages 13 and over. Find every whole number from 1 to 10 million where int(math.log(n, 10)) is wrong. Which numbers are they, and why?
5. Going further. Draw the same power-of-ten bars separately for each funding source. Do the shapes differ?

Guidance of ours. Grades follow age on 31 March, so a class may span nearly a year of birthdays.

More for secundaria pupils is on coding for teenagers in Peru; adults with data at work can look at SQL and data analysis for Peru. All of our Peru guides are gathered on the Peru overview.