Kidus's Chronicles

Today was insanely fun. Such a wonderful memory with Tech nerd Mr. Andy Natiyu It was an amazing day and the movie was something else..we drove a scooter for so long i have hand strapo 😂😂

It even measures from the highest to lowest point on atomic level and what's fascinating is it creates a 3d version of the object we were trying t see.

Touching things at the atomic level (I love microscopes) Adding on the interesting thing i told you guys yesterday, this is a very interesting microscope, this microscope uses touching the specific sample to get image of what it looks like on atomic level.

Good morning, beautiful souls ✨ Growth often feels like loss, because it asks you to let go of who you once were. And if one day you look in the mirror and see someone unfamiliar in a beautiful way, know that you're evolving exactly as you should. Have a peaceful and blessed Thursday 🌿

Okay okay, sooo I can start today. I will tell you guys something that fascinated me after reading and seeing videos about it

Sooo i learn a lot of things, sometimes things i don't think i will use throughout my life, and i love learning new concepts, how things work, the whole shebang, So i was thinking would you guys be interested If i start like one fact a day, i might go deeper and it will be by voice. would you guys be interested? It might be from flowers to atoms to stars to tech, in a wide range of subjects.....

this is a sample dataset where from a ubr data, we anonymized the customers but the same customers have the same name and email. then we can work on many things like anomaly detection, customer behavior, prediction of all sorts

There is one thing to add here, to get consistent identity so that the same fake name/email every time for the same customer or same user You can create a customer identity cache to save the customerId and the same customer in different rows will have the same fake name and email.

def get_consistent_identity(customer_id):
    """ Returns the same fake name/email every time for the same customer id"""
    if customer_id not in customer_identity_cache:
        customer_identity_cache[customer_id] = {
            "full_name": random_name(),
            "email": random_email()
        }
    return customer_identity_cache[customer_id]

this will get consistent identity for the same customer Id that might exist in the record.

Data anonymization is the process of altering, stripping, or replacing personally identifiable information (PII) such as names, emails, phone numbers, and addresses from a dataset so that individual identity cannot be recovered, while still preserving the underlying statistical patterns necessary for Machine Learning (ML) models to learn. In ML workflows, anonymization prevents privacy leaks, ensures regulatory compliance (GDPR, HIPAA, local data protection laws), and prevents models from memorizing specific user identifiers instead of general patterns. Here is a ready-to-publish Telegram post drafted for your channel: Why You MUST Anonymize Data Before Training ML Models (and How to Do It in Python) Focus on "Why" When building Machine Learning models on real-world datasets, you'll often encounter sensitive user details: names, email addresses, locations, or phone numbers. Feeding raw Personally Identifiable Information (PII) into your models is a major risk: 1️⃣ Data Leakage: Models can accidentally memorize and reproduce sensitive training inputs. 2️⃣ Privacy Violations: Violates regulations like GDPR, HIPAA, or local privacy statutes. 3️⃣ Noise: High-cardinality unique IDs add unnecessary noise to training. How Anonymization Works The goal is to scrub or replace identifiers without destroying the distribution or utility of the features. Common strategies include: - Suppression / Stripping: Completely removing columns that offer no predictive value (e.g., social security numbers). - Pseudonymization / Synthetic Replacement: Swapping real identifiers with realistic fake data (e.g., replacing real names/emails with randomized synthetic ones). - Hashing / Tokenization: Mapping sensitive IDs to deterministic hashes or UUIDs when consistent tracking across rows is needed. 🐍 Practical Implementation using Python's Faker Using Python's Faker library, you can quickly generate synthetic replacements for sensitive columns before passing the dataframe to your pipeline:

import pandas as pd
from faker import Faker

fake = Faker()

# Sample dataset with sensitive customer PII
df = pd.DataFrame({
    'name': ['Alice Johnson', 'Bob Smith', 'Charlie Brown'],
    'email': ['alice@gmail.com', 'bob@yahoo.com', 'charlie@outlook.com'],
    'age': [28, 34, 45],
    'purchase_amount': [120.50, 85.00, 310.20]
})

# Anonymize PII columns using synthetic data
df['name'] = [fake.name() for _ in range(len(df))]
df['email'] = [fake.email() for _ in range(len(df))]

print(df)

Result: Your numeric and categorical features (age, purchase_amount) remain intact for the model, while all human identity markers are completely anonymized! 🔒 💡 Pro-Tip: If your downstream tasks require maintaining identical links across tables (e.g., matching historical logs for user X), use deterministic hashing (hashlib) or map original IDs to fake unique hashes (UUID4) using a secure, salt-backed lookup dict! If you have any questions, please feel free to ask...

When shall we have our next live session, Shall we do it this Saturday?

First 4hr Lockin of the day.

Let's all take a break from what we are doing and pray for this kid. I mean I can't stop my mind from thinking the thoughts he might have that day, the plans and everything and suddenly everything went blank, We have no idea how good we got it, Life is incredibly short and fragile and unpredictable. Let's tell our parents we love them and hug them, I can't even think what his mother is thinking right now, I hope he gets fully recovered soon. Donate whatever amount of money you have because i am sure it goes a long way, more than the money, his family will see that he have a swarm of people who would be there. maybe i am venting or so but I can't stop thinking. May God heal him soon,

Gooooood morning beautiful people. Accept what is, let go of what was, have faith in what will be. Have a blessed week ✨

ቂርቆስ እየሉጣን ያወጣ ከሳት እኛንም አድነን አድነን ሊቀ መልእክት እኛንም አድነን አድነን ገብርኤል ሊቀ መልእክት

This is genuinely disheartening. Let's all contribute what we have.

Live. Live. Live