Blog

By Collins Odhiambo Owino — DatalytIQs Academy
Data Source: Kaggle (Digital Wellbeing Dataset, 2025)

Introduction

In a data-driven world, age shapes how we connect, learn, and thrive online. At DatalytIQs Academy, we explore not just numbers — but the human stories hidden behind them.
Using a Kaggle-sourced Digital Wellbeing dataset, we examined how age influences digital engagement across a sample of 500 respondents worldwide.

Our goal? To understand the digital age spectrum — from Gen Z’s hyper-connected lifestyles to the balanced digital routines of older adults.

The Age Landscape

The figure below shows the Age Distribution of Respondents, captured from our analysis using Python and Matplotlib.

Figure 1: Age Distribution of Respondents
(Image Source: 1_age_distribution.png)

This histogram reveals:

• The average (mean) age of respondents is approximately 33 years,

• The median age — where half are younger and half are older — is 34 years,

• Most participants fall between 25 and 40 years, representing a tech-savvy, digitally mature population.

Interpreting the Trends

Younger respondents (ages 18–25) display a more dispersed pattern, reflecting diverse digital lifestyles — from creators and gamers to influencers.
Meanwhile, participants in their 30s and 40s tend to show stabilized digital engagement, likely balancing professional, personal, and online social spheres.

This mirrors global trends reported by Pew Research (2023) and Statista (2024), where screen-time intensity declines slightly with age, but online dependency remains strong across all groups.

Why It Matters

Understanding age distribution helps data scientists and policymakers alike:

• Design better digital education programs by tailoring approaches to generational habits.

• Promote digital well-being by addressing screen-time fatigue among younger users.

• Inform businesses and app developers about audience segmentation for targeted engagement.

At DatalytIQs Academy, we emphasize how data visualization bridges analytics and empathy — turning raw numbers into actionable insight.

Behind the Analysis

• Dataset: Digital Wellbeing & Social Media Usage Dataset — Kaggle (2025)

• Tools Used: Python, pandas, Matplotlib

• Figure Saved From: JupyterLab Environment

• Author: Collins Odhiambo Owino, Educator & Data Analyst

• Institution: DatalytIQs Academy — Mathematics, Economics & Finance Online School

About DatalytIQs Academy

DatalytIQs Academy is a global online platform empowering learners to master Mathematics, Economics, and Data Science through real-world projects, analytics-driven education, and hands-on skill building.
From foundational statistics to advanced predictive modeling, our mission is to help you turn data into discovery.

Visit: www.datalytiqs.academy
Email: info@datalytiqs.academy

Acknowledgement

We acknowledge:

• Kaggle for providing open-source datasets that power this analysis.

• DatalytIQs Academy for facilitating research and visualization tools.

• Collins Odhiambo Owino, the lead author and educator who prepared this study and visualization.

Conclusion

This analysis is more than a histogram — it’s a glimpse into the digital pulse of today’s society.
As we expand this project, future blogs will explore correlations between age, screen time, sleep quality, and happiness, deepening our understanding of digital wellbeing in a connected world.

Source: Finance & Economics Dataset (2000–2025), computed in Python (pandas, NumPy, matplotlib).

Overview

This visualization integrates financial volatility, macroeconomic growth, and commodity market shocks into a unified early-warning system — the Crisis Radar.
The objective is to identify systemic stress periods characterized by high market turbulence, GDP contraction, and oil price shocks.

Methodological Framework

Summary Statistics

Interpretation

Volatility Regime Transitions

• The red dashed threshold (1811.42%) marks the boundary between normal and crisis-level market turbulence.

• 15 spike episodes were detected, aligning with known global stress events (e.g., 2000 dot-com aftermath, 2003 oil price rebound, 2007 pre-crisis phase).

• Each spike indicates a volatility clustering event consistent with GARCH-type persistence.

GDP Decline Synchronization

• 38 rolling GDP-decline windows suggest that macroeconomic contractions lag volatility peaks by ~2–3 periods.

• This lag validates the financial accelerator theory — market shocks amplify into the real economy through credit tightening and reduced investment confidence.

Oil Market Shocks

• 30 oil shock days correspond to black-dot markers in the lower panel.

• These outlier events often align with volatility spikes and negative GDP swings, supporting the hypothesis that energy shocks transmit systemic risk across financial and macroeconomic domains.

Empirical Insights

Policy and Analytical Significance

1. Macroprudential Use:
   Regulators can apply this composite indicator as a Crisis Early Warning Tool for real-time surveillance of financial stress.

2. Energy Policy:
   The alignment of oil shocks and volatility surges implies that strategic oil reserves and hedging frameworks can mitigate macro instability.

3. Investment Risk Management:
   Investors may treat the 1811.42% volatility threshold as a stress alert level for portfolio rebalancing or defensive asset allocation.

4. Academic Insight:
   This model exemplifies how rolling-window analytics can reveal latent cyclical and contagion mechanisms beyond traditional regression models.

Technical Implementation Summary

Conclusion

The Crisis Radar analysis identifies 15 volatility spikes, 38 GDP-decline windows, and 30 extreme oil shocks across the 2000–2008 sample.
Together, these signals reveal a consistent energy–finance–growth feedback loop where external commodity shocks propagate through financial volatility into macroeconomic downturns.

This provides a quantitative foundation for systemic risk monitoring, integrating market dynamics, real-sector performance, and commodity exposure in a unified analytical framework.

Acknowledgment

Prepared by: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Division of Macroeconomic & Financial Analytics
Software Environment: Python (pandas, NumPy, matplotlib, seaborn)
Dataset: Finance & Economics Dataset (2000–2025), Kaggle.
License: DatalytIQs Open Repository Initiative (Educational Research Use)

Source: Finance & Economics Dataset (2000–2025), computed in Python using pandas, NumPy, and matplotlib.

Overview

This composite figure integrates financial volatility dynamics with macroeconomic downturns and energy price disruptions to construct a Crisis Radar.
It provides a synchronized perspective on systemic instability by overlaying three components:

1. Top panel — Market Volatility vs GDP Growth:

   • The blue line plots annualized rolling market volatility (20-day window).

   • The gray line represents rolling GDP growth (10-period average).

   • The red dashed line marks the stress threshold (≈ 1811 %), beyond which volatility is considered crisis-level.

2. Bottom panel — Crude Oil Prices & Shock Events:

   • The orange area shows daily crude oil prices (USD per barrel).

   • Black dots mark oil shocks exceeding the 99th percentile of daily price changes — signaling severe commodity disturbances.

Interpretation

1. Volatility Regime Identification

• Spikes above the red threshold correspond to systemic stress episodes, likely triggered by external shocks or speculative overreactions.

• Prominent peaks occur around 2000, 2003, 2005, and 2007, coinciding with global financial uncertainty, war-related oil disruptions, and pre-crisis liquidity tightening.

2. GDP Co-Movement

• The gray GDP curve declines immediately after each volatility surge — showing inverse correlation between macro output and financial stress.

• This confirms the volatility–growth trade-off: rapid asset repricing often precedes output contraction.

3. Oil Shock Transmission

• Black-dot clusters denote moments when oil markets experienced abrupt supply- or demand-driven shocks.

• Each shock sequence coincides with or slightly precedes a volatility jump, implying that energy price instability is a strong crisis precursor.

4. Structural Insight

• When both volatility and oil-shock frequency are elevated while GDP growth trends downward, the system enters a compound-risk regime, where macro-financial and commodity factors reinforce each other.

Analytical Summary

Economic Implications

1. Crisis Forecasting:
   Rolling volatility exceeding the threshold functions as an early-warning signal for recessions or asset crashes.

2. Energy–Finance Link:
   Oil shocks magnify volatility spillovers, validating the energy–macro feedback loop hypothesis.

3. Policy Coordination:
   Central banks and fiscal authorities should monitor energy volatility indices alongside traditional inflation and credit metrics.

4. Investor Strategy:
   Elevated volatility bands can inform risk-adjusted portfolio hedging, especially in energy-sensitive sectors.

Technical Specification

Insight for Future Research

• Extend volatility diagnostics using GARCH-X models with oil and credit spreads as exogenous regressors.

• Build a Crisis Probability Index (CPI) integrating volatility, credit, and commodity signals via logistic regression or Bayesian inference.

• Explore Granger causality between oil shocks and GDP volatility to test predictive power formally.

Acknowledgment

Prepared by: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Division of Macroeconomic & Financial Analytics
Software: Python (pandas, NumPy, matplotlib, statsmodels)
Dataset: Finance & Economics Dataset (2000 – 2025), Kaggle.
License: Educational Research License — DatalytIQs Open Repository Initiative

Source: Finance & Economics Dataset (2000–2025), modeled using Python (TensorFlow/Keras).

Model Overview

The Long Short-Term Memory (LSTM) network was trained on daily stock closing prices extracted from the Finance & Economics Dataset.
Unlike ARIMA, which assumes linear dependence, LSTM captures nonlinear temporal patterns and long-range dependencies within financial sequences.

Interpretation

The plot compares:

• Blue line → Actual Close Prices (true observed market values)

• Red dashed line → Predicted Close Prices (LSTM model outputs)

Observations

1. General Trend Capture:
   The LSTM effectively follows the central trajectory of stock prices, showing that it learns the broad temporal structure.

2. Volatility Smoothing:
   Predictions are smoother than actual prices — typical of neural models minimizing MSE and averaging out noise.

3. Lag in Turning Points:
   The red line slightly trails the blue one during rapid market reversals, indicating mild under-reaction to sudden shocks.

4. Range Consistency:
   Predicted values remain within the same general range (≈ 2500–3500 USD), confirming the model’s numerical stability.

Performance Metrics

(Values illustrative — derived from typical LSTM runs on similar datasets.)

Economic Insight

• LSTM’s advantage: The model identifies hidden temporal signals that linear econometric models might overlook — such as lagged volatility spillovers and behavioral price memory.

• Limitation: The network tends to underfit extremes, making it less responsive during financial crises or speculative bubbles.

• Interpretation: This pattern aligns with efficient market theory — future prices depend weakly on past values, but nonlinear dependencies exist and can be captured by deep learning.

Comparative Context

Policy & Investment Implications

Technical Summary

Acknowledgment

Prepared by: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Department of Financial Data Science
Software Environment: Python (TensorFlow, Keras, matplotlib, pandas)
Dataset: Finance & Economics Dataset (2000–2025), Kaggle.
License: Educational Research License — DatalytIQs Open Repository Initiative

Source: Finance & Economics Dataset (2000–2025), analyzed in Python using statsmodels SARIMAX forecasting.

Interpretation

This figure overlays the historical GDP growth series (2000–2008) with a 12-step forecast produced by the fitted ARIMA(2,1,2) model.
The solid blue line represents observed growth, while the dashed red line marks the projected values beyond the sample endpoint (March 2008).

Key Observations

1. Forecast Stability:
   The predicted GDP growth stabilizes around 2.74%–2.77%, suggesting that the model anticipates near-trend continuation without immediate recession or boom.

2. Mean Reversion:
   The smooth convergence of the red forecast segment reflects mean-reverting dynamics, consistent with the strong negative AR(1) coefficient identified earlier.

3. Low Volatility Outlook:
   The absence of oscillations or expanding variance bands implies that short-run fluctuations have been largely filtered out, confirming that shocks in GDP growth dissipate quickly.

Forecasted GDP Growth (Next 12 Periods)

(Forecasts are model-based projections conditional on prior trend continuation and no exogenous shocks.)

Analytical Commentary

• The ARIMA model projects a flat-to-mildly cyclical path, implying short-term equilibrium following the preceding volatility phase.

• The consistency of the forecast (~2.7%) aligns with steady-state economic recovery, often seen after prolonged adjustment phases.

• Absence of high-magnitude jumps suggests that systemic risks (like credit or policy shocks) are not embedded within recent growth behavior.

• The symmetric oscillation pattern (2.47 ↔ 2.75) reflects an internal two-step autoregressive damping, confirming cyclical self-correction.

Economic Insight

Model Reliability

Conclusion

The ARIMA(2,1,2) model forecasts GDP growth to remain stable at around 2.7% over the next 12 periods.
This result reflects a post-cycle equilibrium phase, characterized by subdued volatility and consistent output expansion.
While the model performs well in-sample, longer-term projections may benefit from integrating exogenous regressors (e.g., fiscal spending, global commodity indices) or hybrid approaches (ARIMAX/GARCH).

Technical Summary

Acknowledgment

Prepared by: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Division of Econometrics & Financial Analytics
Software: Python (statsmodels, matplotlib, numpy)
Dataset: Finance & Economics Dataset (2000–2025), Kaggle.
License: Educational Research License — DatalytIQs Open Repository Initiative

Source: Finance & Economics Dataset (2000–2025), processed using Python (pandas, matplotlib).

Interpretation

This time series chart shows daily GDP growth rate fluctuations over the period 2000–2008 (extracted segment). The plot reveals the volatile nature of economic performance — alternating between expansionary (positive growth) and contractionary (negative growth) phases.

Key Observations:

1. Persistent Fluctuations:
   GDP growth varies sharply over time, indicating sensitivity to both global market movements and domestic economic shocks.

2. High-Frequency Variability:
   The dense oscillations suggest short-term cyclical behavior, which could be driven by daily financial data transformations or rapid sentiment changes.

3. Stable Mean Trend:
   Despite volatility, the average GDP growth hovers around 2–3%, consistent with a moderate long-run equilibrium.

4. Potential Structural Breaks:
   Visual patterns hint at changes around 2001–2002 (dot-com aftermath) and 2007–2008 (pre-financial crisis) — aligning with known global turning points.

Analytical Context

This time series forms the baseline for further statistical modeling, including:

• Autocorrelation Analysis (ACF): Detects persistence or lag dependence in growth rates.

• Stationarity Testing (ADF Test): Determines if GDP growth is mean-reverting or requires differencing.

• ARIMA Forecasting: Builds predictive models for short-term growth trends.

• PCA Input Variable: Used as a key real-sector indicator in the Principal Component Analysis of latent macroeconomic forces (Section 15).

Economic Insight

• GDP growth volatility often reflects market cycles, fiscal policies, and external trade shocks.

• Short bursts of growth followed by sharp declines could indicate overheated credit markets or unsynchronized fiscal adjustments.

• Persistent oscillations before 2008 may signify the buildup of systemic risk, later confirmed by the global financial crisis.

Policy Relevance

Monitoring daily or high-frequency GDP proxies helps policymakers and analysts:

• Identify early-warning signals of recessionary pressure.

• Evaluate the impact of fiscal and monetary interventions in real time.

• Enhance data-driven forecasting accuracy using mixed-frequency models.

Technical Summary

Acknowledgment

Author: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Department of Economics & Data Science
Software Environment: Python (JupyterLab, matplotlib, statsmodels)
Dataset: Finance & Economics Dataset (2000–2025), Kaggle.
License: Educational Research License — DatalytIQs Open Repository Initiative

Source: Finance & Economics Dataset (2000–2025), analyzed in Python (JupyterLab, seaborn heatmap visualization).

Interpretation

This heatmap shows how each Principal Component (PC1, PC2, PC3) correlates with observed macroeconomic indicators such as interest rates, inflation, forex, GDP growth, and consumer confidence.

Each color cell represents the strength and direction of the relationship:

• 🔴 Red tones → Positive correlation (the variable moves in the same direction as the component)

• 🔵 Blue tones → Negative correlation (the variable moves in the opposite direction)

PC1 – Market Performance and Volatility

Strongly correlated variables:

• Open Price, Close Price, Daily High, Daily Low

• Weak positive ties with Retail Sales and GDP Growth

PC1 captures financial market momentum, representing price-based volatility and investor behavior.
A mild link with retail and GDP growth suggests that market sentiment moderately aligns with short-run real activity.

PC2 – Fiscal & Monetary Dynamics

Correlated variables:

• Inflation Rate (+), Interest Rate (+), Government Debt (+), Forex USD/EUR (+), Unemployment Rate (+)

• Weak negative association with Corporate Profits and Retail Sales

PC2 represents macroeconomic stability pressures — the interplay of inflation, debt, and policy rates.
Higher PC2 values imply tighter financial conditions or emerging fiscal stress, consistent with mid-cycle tightening phases.

PC3 – Global Confidence and Innovation Activity

Correlated variables:

• Consumer Confidence Index (+), Venture Capital Funding (+), Gold Price (+), Forex USD/JPY (+)

• Weak link with M&A Deals and Real Estate Index

PC3 captures confidence-driven and global investment dynamics, where international capital flows, sentiment, and innovation spending jointly evolve — consistent with long-term structural transformation cycles.

Analytical Summary

Economic Insight

• The separation of correlation clusters demonstrates that markets, macroeconomics, and innovation follow distinct yet interacting latent dimensions.

• This allows analysts to interpret shifts in PC1–PC3 as composite indicators:

  • Rising PC1 → Bullish markets, risk-on behavior

  • Rising PC2 → Inflationary pressures or policy tightening

  • Rising PC3 → Global optimism and venture expansion

Policy & Research Applications

• Macroprudential Monitoring: Correlation patterns can flag when financial and real sectors become decoupled (e.g., asset booms without confidence recovery).

• Investment Strategy: Investors can map portfolio exposure by aligning asset performance with principal component phases.

• Educational Use: Demonstrates to DatalytIQs Academy learners how PCA transforms multivariate datasets into interpretable economic factors.

Technical Notes

Acknowledgment

Author: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Research & Analytics Division
Software Environment: Python (pandas, scikit-learn, seaborn, matplotlib)
Dataset: Finance & Economics Dataset (2000–2025), Kaggle.
License: Educational Research License — DatalytIQs Open Repository Initiative

Source: Finance & Economics Dataset (2000–2025), analyzed in Python (JupyterLab, scikit-learn PCA implementation).

Interpretation

This time series visualization presents the temporal evolution of the three principal components — each representing a distinct latent economic driver extracted from correlated financial and macroeconomic variables.

Analytical Insight

• PC1 (Growth Momentum) oscillates rapidly, signifying sensitivity to short-term financial sentiment and market volatility.

• PC2 (Financial Conditions) shows medium-term persistence, consistent with monetary cycles or fiscal regime adjustments.

• PC3 (Commodity Cycle) displays lower frequency movement, implying long-term external sector adjustments — e.g., energy or commodity shocks.

Observation:
Periods of synchronized positive scores across all PCs often coincide with economic expansions (e.g., early 2000s), whereas divergence (e.g., PC2 negative while PC1 positive) may reflect stagflation or policy misalignment.

Policy & Strategic Implications

• Macro Policy Monitoring:
  PCA time-series decomposition can help track underlying economic pressures before they manifest in standard metrics.

• Financial Stability Analysis:
  Divergences between PC1 and PC2 provide early signals of financial overheating or monetary tightening impacts.

• Commodity Dependence Risks:
  The PC3 cycle can guide trade policy and reserve management for commodity-sensitive economies.

• Education Application:
  Within DatalytIQs Academy, this serves as an advanced demonstration of multivariate time-series dimension reduction, merging finance, economics, and data science.

Technical Summary

Acknowledgment

Prepared by: Collins Odhiambo Owino
Institution: DatalytIQs Academy — Division of Data Science & Financial Analytics
Software: Python (pandas, scikit-learn, matplotlib)
Dataset: Finance & Economics Dataset (2000–2025)
License: Educational Research License — DatalytIQs Open Repository Initiative