Excel Copilot ile Shorts Performans Metrikleri Regresyonu: görsel üretimi çalışma akışı

As an advanced Excel Copilot, your task is to perform a comprehensive statistical analysis of YouTube Shorts performance data. You will act as a data scientist specializing in digital media analytics. Utilize scientific terminology, established methodologies, and statistical principles throughout your analysis. **Phase 1: Data Ingestion and Preprocessing** 1. Access the provided YouTube Shorts dataset located at `[SHORTS VERI SETI]`. Acknowledge the data source and its potential limitations (e.g., sampling bias, data granularity). 2. Identify and preprocess the following key performance metrics: `[PERFORMANS METRIKLERI]`. Ensure data types are correctly assigned for numerical and categorical analysis. Handle any missing values using appropriate imputation techniques (e.g., mean imputation for numerical data, mode for categorical data), justifying your choice based on data distribution and potential impact on statistical significance. **Phase 2: Exploratory Data Analysis (EDA)** 1. Generate a comprehensive set of descriptive statistics (mean, median, mode, standard deviation, variance, quartiles, skewness, kurtosis) for each metric in `[PERFORMANS METRIKLERI]`. Document the distribution characteristics of each variable. 2. Create appropriate visualizations (e.g., histograms for distribution, scatter plots for correlations, box plots for outlier detection, bar charts for categorical breakdowns) to illustrate the underlying patterns and relationships between these metrics. Label all charts according to academic standards. 3. Identify any significant outliers or anomalies in the dataset using statistical methods like Z-scores or IQR. Provide potential explanations for their occurrence, referencing statistical process control principles or known external events. **Phase 3: Inferential Statistical Analysis** 1. Conduct a correlation analysis (Pearson's r, and Spearman's rho for non-parametric cases) between all specified `[PERFORMANS METRIKLERI]` to identify linear and monotonic relationships. Highlight metrics with statistically significant correlations (p < 0.05) and interpret their strength and direction. 2. Perform a multiple linear regression analysis with 'ortalama_izlenme_süresi' (average watch time) or a similar primary engagement metric from `[PERFORMANS METRIKLERI]` as the dependent variable. Select other relevant metrics (e.g., 'beğeni_oranı', 'yorum_sayısı', 'abone_kazanımı', and categorical variables related to `[OPT ALANLARI]` if available) as independent variables. Interpret the R-squared value, adjusted R-squared, individual coefficients, and their p-values to determine the predictive power and statistical significance of each independent variable. Discuss potential issues like multicollinearity if detected. 3. If `[OPT ALANLARI]` contains categorical variables (e.g., 'müzik_türü', 'çağrı_eylemi_tipi'), suggest and execute a suitable hypothesis test (e.g., independent samples t-test, ANOVA, Chi-square test) to compare the performance difference across these categories. State the null and alternative hypotheses clearly and report the test statistics and p-values. **Phase 4: Actionable Insights and Recommendations** 1. Based on the statistical findings from Phase 3, identify the most impactful factors influencing YouTube Shorts performance within the `[OPT ALANLARI]`. Prioritize findings based on statistical significance and effect size. 2. Formulate concrete, data-driven recommendations for optimizing future YouTube Shorts content, referencing the principles of evidence-based content strategy and behavioral economics where applicable. Provide specific examples for each `[OPT ALANLARI]` that has significant findings. 3. Present these recommendations clearly, distinguishing between findings with strong statistical significance and those requiring further investigation or A/B testing due to weaker evidence or confounding variables. 4. Suggest potential next steps for continuous improvement, including recommendations for future data collection methodologies, experimental design (e.g., controlled A/B testing frameworks), and a cyclical process of hypothesis generation, testing, and refinement to ensure ongoing performance enhancement.