grammarly typos

app.py

@@ -12,12 +12,12 @@ you may be aware that changes in the distribution of
 the production data can affect the model's performance.
 """)
             
-st.markdown("""Recently a paper from MIT, Harvard and other institutions showed how [91% of their ML models
-experiments degradated]('https://www.nannyml.com/blog/91-of-ml-perfomance-degrade-in-time') in time.""")
+st.markdown("""Recently a paper from MIT, Harvard, and other institutions showed how [91% of their ML models
+experiments degraded]('https://www.nannyml.com/blog/91-of-ml-perfomance-degrade-in-time') in time.""")
             
-st.markdown("""Typically, to know if a model is degrading we need access ground truth. But most of the times
-getting new labeled data is either expensive, takes lots of time or imposible. So we end up blindless without
-knowing how the model is performing in production. 
+st.markdown("""Typically, we need access to ground truth to know if a model is degrading.
+But most of the time, getting new labeled data is expensive, time-consuming, or impossible.
+So we end up blindless without knowing how the model performs in production. 
 """)
 
 st.markdown("""
@@ -39,18 +39,18 @@ car_value, salary_range, loan_lenght, etc.
 st.dataframe(analysis_df.head(3))
 
 st.markdown("""
-We know that the model had a **Test F1-Score of: 0.943**. But, what guarantees us that the F1-Score
+We know that the model had a **Test F1-Score of: 0.943**. But what guarantees us that the F1-Score
 will continue to be good on production data?
 """)
 
 st.markdown("#### Estimating the Model Performance")
 st.markdown("""
-Instead of waiting for ground truth we can use NannyML's 
+Instead of waiting for ground truth, we can use NannyML's 
 [CBPE]("https://nannyml.readthedocs.io/en/stable/tutorials/performance_estimation/binary_performance_estimation/standard_metric_estimation.html")
 method to estimate the performance of an ML model.
 
 CBPE's trick is to use the confidence scores of the ML model. It calibrates the scores to turn them into actual probabilities.
-Once the probabilities are calibrate it can estimate any performance metric that can be computed from the confusion matrix elements.
+Once the probabilities are calibrated, it can estimate any performance metric that can be computed from the confusion matrix elements.
             """)
 
 chunk_size = st.slider('Chunk/Sample Size', 2500, 7500, 5000, 500)
@@ -101,7 +101,7 @@ st.divider()
 
 
 
-st.markdown("""Created by [santiviquez](https://twitter.com/santiviquez) from NannyML""")
+st.markdown("""Created by [santiviquez](https://twitter.com/santiviquez) from NannyML.""")
 
 st.markdown("""
 NannyML is an open-source library for post-deployment data science. Leave us a 🌟 on [GitHub]("https://github.com/NannyML/nannyml")