Online coding and AI classes in the Loreto region

INEI, Censos Nacionales 2025, table 1 of the demographic indicators: people counted plus INEI's estimate of those missed, on 15 September 2025, by the boundaries in force on 15 July 2026. Rows printed as published; the province figures are not ours to add up.

From cuadros 5.16, 5.17, 5.22 and 5.23 of INEI's Compendio Estadístico Perú 2026, which reproduce MINEDU's school census. We swap the decimal comma for a point and add nothing up.

At national level, MINEDU's Python programme, Fábrica de Programadores, started in July 2025 with 1,000 pupils in the last two years of secundaria in Lima Norte. In September 2026 the ministry announced a plan to bring satellite internet to 55,000 public schools. Both are steps; neither puts a programming teacher in a Loreto classroom this year.

1. Book the free class at your usual hour. Choose the time you would really use, so the test reflects normal conditions in your home.
2. Start with sound, not video. The mentor's voice and the shared screen matter most. Cameras can stay off if the picture freezes.
3. Share code as text. Programs can be pasted into the chat, so nobody needs to read code from a blurred video frame.
4. Keep one tab open. Streaming or downloads elsewhere in the house slow the class; pausing them during the hour helps.
5. Decide together. After the class the mentor tells you honestly whether it worked well enough, and what could change it, before you pay anything.

If a lesson drops, the mentor waits and continues where the learner left off. In one-to-one plans a missed slot can be moved. This is not a promise that every connection will do; it is a promise to tell you the truth after the test.

For a national overview of how we teach Peruvian learners in English, see coding and AI classes in Peru. Parents of younger children may also like Python for kids in Peru, which explains when a child is ready to type.

import csv
import re

PH = re.compile(r"pH (\d\.\d+)")

values = []
with open("horizontes_suelo.csv", newline="", encoding="utf-8-sig") as f:
    for row in csv.DictReader(f):
        if row["id_perfil"].startswith("LO-"):
            found = PH.search(row["descripcion_campo"])
            if found:
                values.append(float(found.group(1)))

values.sort()
print("Loreto layers with a pH:", len(values))
print("Lowest:", values[0], "Highest:", values[-1])
print("Middle value:", values[len(values) // 2])

The pattern pH followed by a space, a digit, a point and more digits found a value in every one of the 145 layers. The middle value, 4.19, is strongly acidic soil, typical of the old, heavily rained-on ground of the lowland Amazon. A learner could stop here. But the text holds a second description of the same fact, the word label, and two descriptions of one fact should agree.

Version 2 captures the words just before the bracket as well, groups the layers by label, and prints the lowest and highest pH under each one. If the labels are consistent, each range should sit neatly above the one before it.

import csv
import re

LABELLED = re.compile(r"([a-záéíóúñ ]+?)\s*\(pH (\d\.\d+)")

ranges = {}
with open("horizontes_suelo.csv", newline="", encoding="utf-8-sig") as f:
    for row in csv.DictReader(f):
        if not row["id_perfil"].startswith("LO-"):
            continue
        found = LABELLED.search(row["descripcion_campo"])
        if found:
            label = found.group(1).strip().lower()
            ranges.setdefault(label, []).append(float(found.group(2)))

for label, ph in sorted(ranges.items(), key=lambda item: min(item[1])):
    print(f"{label:24} {len(ph):3} layers, pH {min(ph)} to {max(ph)}")

Three things jump out. First, a typing slip: two layers say fultra ácido, which is ultra ácido with a stray letter, and a program treats them as a separate category until someone fixes it. Second, label variants: extremadamente ácida with an a, and reacción fuertemente ácida with an extra word. Third, an overlap: the ultra ácido range reaches 4.59, higher than every extremadamente ácido layer, even though ultra is meant to be the more acidic word. Checking that row shows a single layer, horizon 3 of profile LO-029, labelled ultra ácido at pH 4.59; every other ultra ácido layer is below 3.6.

We do not correct the IIAP's records. We report what the program found and let a soil scientist decide whether the word or the number is wrong. That is the habit we want every learner to leave with: a pattern that matches is not the same as a fact that is right.

Data credit: IIAP, soil profiles of San Martín, Loreto, Madre de Dios and Ucayali, reused under ODC-By. The summaries are ours. One layer, in profile LO-013, opens a bracket before its pH and never closes it, so a stricter pattern that demanded the closing bracket would miss it.

1. Under 8, in the garden. Sort five words, warm, hot, boiling, cool, cold, into order. Then ask: what if someone wrote "boiling" for a cup of tea that was only warm?
2. Block coding, around 8 to 10. Make a list of pH numbers and a loop that says "acid" below 7, "neutral" at 7 and "alkaline" above it.
3. First Python, around 10 to 12. Run Version 1, then change the pattern so it insists on the closing bracket, and find the layer it now misses.
4. Ages 13 and over. Write a dictionary that merges fultra ácido into ultra ácido and the feminine forms into the masculine ones, then print the ranges again.
5. Going further. List every pair of labels whose ranges overlap, and the layers responsible. Write two sentences to the IIAP describing what you found, without claiming which value is wrong.

Course ages are ours. MINEDU counts normative age on 31 March, so a pupil with a birthday after that date may be one grade below the row.

University students in Iquitos can read about coding courses for university students in Peru, and adults returning to learning later in life have their own page on coding after 50 in Peru.

Per the tz database, America/Lima is UTC minus five hours with no daylight saving since 1994, and INACAL keeps Peru's legal time. India is UTC plus five and a half hours. Iquitos and India are therefore ten and a half hours apart all year.

No clocks change on either side, so a slot that works in the rainy season works in the drier months too. Early mornings in Iquitos fall in the Indian evening, which suits many mentors, and they can also be the calmest time on a shared home connection.