Implement basic statistics

codestrange · Dec 6, 2020 · d591cb1 · d591cb1
1 parent 9cf51d8
commit d591cb1
Show file tree

Hide file tree

Showing 8 changed files with 191 additions and 16 deletions.
diff --git a/app.py b/app.py
@@ -1,13 +1,19 @@
 from typing import Dict, List, Tuple
 
 import pandas as pd
+import plotly.graph_objects as go
 import streamlit as st
 from PIL import Image
 
 from kickstarter.core import get_favorite_categories as get_favorite_categories_raw
 from kickstarter.core import load_json as load_json_raw
 from kickstarter.models import CategoryModel, ProjectModel
-from kickstarter.processing import GrossingCategoriesModel, SuccessfulCategoriesModel
+from kickstarter.processing import (
+    GrossingCategoriesModel,
+    MonthlyCategoriesSuccessModel,
+    MonthlyCategoriesTotalsModel,
+    SuccessfulCategoriesModel,
+)
 
 image = Image.open("images/logo.png")
 st.image(image, use_column_width=True)
@@ -29,6 +35,9 @@ def get_favorite_categories(
     List[CategoryModel],
     List[CategoryModel],
     List[CategoryModel],
+    MonthlyCategoriesSuccessModel,
+    MonthlyCategoriesTotalsModel,
+    Dict[int, Tuple[List[int], List[int]]],
 ]:
     return get_favorite_categories_raw(projects, categories)
 
@@ -98,6 +107,9 @@ def get_favorite_categories(
     List[CategoryModel],
     List[CategoryModel],
     List[CategoryModel],
+    MonthlyCategoriesSuccessModel,
+    MonthlyCategoriesTotalsModel,
+    Dict[int, Tuple[List[int], List[int]]],
 ] = get_favorite_categories(
     *load_json()  # type: ignore
 )
@@ -167,9 +179,43 @@ def get_favorite_categories(
 
 En las siguientes gráficas mostramos la cantidad de proyectos, asi como la
 cantidad de estos que fueron exitosos, de estas categorías por mes durante
-los años del 2009 al 2018.
+los años del 2009 al 2020.
 """
 
+timeline = go.Figure()  # type: ignore
+
+for cat in favorite_categories[4]:
+    timeline.add_scatter(
+        x=[
+            favorite_categories[6].dates[i]
+            for i in range(len(favorite_categories[5].dates))
+        ],
+        y=[
+            favorite_categories[6].categories[cat.id][i]
+            for i in range(len(favorite_categories[5].dates))
+        ],
+        name=cat.name,
+        opacity=0.9,
+    )
+    timeline.add_scatter(
+        x=[
+            favorite_categories[5].dates[i]
+            for i in range(len(favorite_categories[5].dates))
+        ],
+        y=[
+            favorite_categories[5].categories[cat.id][i]
+            for i in range(len(favorite_categories[5].dates))
+        ],
+        name=cat.name + " Exitosos",
+        opacity=0.9,
+    )
+
+timeline.update_layout(
+    title_text="Cantidad de Proyectos vs Cantidad de Proyectos Exitosos",
+    xaxis_rangeslider_visible=True,
+)
+st.write(timeline)
+
 """
 Teniendo en cuenta la información anterior podemos notar que muchas de las
 categorías anteriores tuvieron un auge en **Kickstarter** durante los años
@@ -193,3 +239,42 @@ def get_favorite_categories(
 A continuación analizaremos como se comportan las categorías que han sido y son
 más exitosas dependiendo del més del año en que sus proyectos fueron dados a conocer.
 """
+
+category_selected = st.selectbox(
+    "Categorías:",
+    options=[item.id for item in favorite_categories[4]],
+    format_func=lambda x: favorite_categories[0].categories[x].name,
+)
+
+months = [
+    "enero",
+    "febrero",
+    "marzo",
+    "abril",
+    "mayo",
+    "junio",
+    "julio",
+    "agosto",
+    "septiembre",
+    "obtubre",
+    "noviembre",
+    "diciembre",
+]
+
+successfuls = favorite_categories[7][category_selected][0]
+totals = favorite_categories[7][category_selected][1]
+
+fig_cat_months_rel = go.Figure(  # type: ignore
+    data=[
+        go.Bar(name="Total", x=months, y=totals),  # type: ignore
+        go.Bar(name="Éxitos", x=months, y=successfuls),  # type: ignore
+    ]
+)
+
+fig_cat_months_rel.update_layout(
+    title_text="Categoría: "
+    + f"{favorite_categories[0].categories[category_selected].name}",
+    barmode="overlay",
+)
+
+st.write(fig_cat_months_rel)
diff --git a/kickstarter/core.py b/kickstarter/core.py
@@ -5,7 +5,13 @@
 from .models import CategoryModel, ProjectModel
 from .processing import top_grossing_categories  # noqa: F401
 from .processing import top_successful_categories  # noqa: F401
-from .processing import GrossingCategoriesModel, SuccessfulCategoriesModel, process
+from .processing import (
+    GrossingCategoriesModel,
+    MonthlyCategoriesSuccessModel,
+    MonthlyCategoriesTotalsModel,
+    SuccessfulCategoriesModel,
+    process,
+)
 
 
 def load_json() -> Tuple[
@@ -36,12 +42,19 @@ def get_favorite_categories(
     List[CategoryModel],
     List[CategoryModel],
     List[CategoryModel],
+    MonthlyCategoriesSuccessModel,
+    MonthlyCategoriesTotalsModel,
+    Dict[int, Tuple[List[int], List[int]]],
 ]:
     results = process(projects, categories)
     grossing_categories: List[CategoryModel] = []
     successful_categories: List[CategoryModel] = []
     crossing_categories_model: GrossingCategoriesModel = None  # type: ignore
     successful_categories_model: SuccessfulCategoriesModel = None  # type: ignore
+    monthly_categories_success_model: MonthlyCategoriesSuccessModel = (
+        None
+    )  # type: ignore
+    monthly_categories_totals_model: MonthlyCategoriesTotalsModel = None  # type: ignore
     for item in results:
         if isinstance(item, GrossingCategoriesModel):
             item = cast(GrossingCategoriesModel, item)
@@ -51,16 +64,37 @@ def get_favorite_categories(
             item = cast(SuccessfulCategoriesModel, item)
             successful_categories = item.top[:25]
             successful_categories_model = item
+        elif isinstance(item, MonthlyCategoriesSuccessModel):
+            item = cast(MonthlyCategoriesSuccessModel, item)
+            monthly_categories_success_model = item
+        elif isinstance(item, MonthlyCategoriesTotalsModel):
+            item = cast(MonthlyCategoriesTotalsModel, item)
+            monthly_categories_totals_model = item
     grossing_categories_set = set(grossing_categories)
     successful_categories_set = set(successful_categories)
     inter = cast(
         List[CategoryModel],
         list(set.intersection(grossing_categories_set, successful_categories_set)),
     )
+    by_months: Dict[int, Tuple[List[int], List[int]]] = {}
+    for item in categories:
+        by_months[item] = ([0] * 12, [0] * 12)
+    dates = monthly_categories_success_model.dates
+    for item in by_months:
+        for (index, success), (_, total) in zip(
+            monthly_categories_success_model.categories[item].items(),
+            monthly_categories_totals_model.categories[item].items(),
+        ):
+            date = dates[index]
+            by_months[item][0][date.month - 1] += success
+            by_months[item][1][date.month - 1] += total
     return (
         crossing_categories_model,
         successful_categories_model,
         grossing_categories,
         successful_categories,
         list(inter),
+        monthly_categories_success_model,
+        monthly_categories_totals_model,
+        by_months,
     )
diff --git a/kickstarter/processing/__init__.py b/kickstarter/processing/__init__.py
@@ -1,4 +1,12 @@
 from .base_process import process, subscribe  # noqa: F401
+from .monthly_categories_success import (  # noqa: F401
+    MonthlyCategoriesSuccessModel,
+    monthly_categories_success,
+)
+from .monthly_categories_total import (  # noqa: F401
+    MonthlyCategoriesTotalsModel,
+    monthly_categories_totals,
+)
 from .top_grossing_categories import (  # noqa: F401
     GrossingCategoriesModel,
     top_grossing_categories,

diff --git a/kickstarter/processing/monthly_categories_success.py b/kickstarter/processing/monthly_categories_success.py
@@ -1,6 +1,6 @@
 from collections import defaultdict
 from datetime import datetime
-from typing import Dict, List, Optional, Tuple
+from typing import Dict, Optional
 
 from ..models import CategoryModel, ProjectModel
 from . import subscribe
@@ -11,20 +11,25 @@ def __init__(self, start_year=2009, end_year=2020):
         counter: int = 0
         self.dates: Dict[int, datetime] = dict()
         self._index: Dict[datetime, int] = dict()
-        self.categories: Dict[int, Dict[int, int]] = defaultdict(lambda: defaultdict(lambda: 0))
+        self.categories: Dict[int, Dict[int, int]] = defaultdict(
+            lambda: defaultdict(lambda: 0)
+        )
 
-        for year in range(start_year, end_year+1):
+        for year in range(start_year, end_year + 1):
             for month in range(1, 13):
                 date = datetime(year, month, 1)
                 self.dates[counter] = date
                 self._index[date] = counter
                 counter += 1
 
     def count(self, project: ProjectModel):
-        to_count_date: datetime = datetime(project.state_changed_at.year, project.state_changed_at.month, 1)
+        to_count_date: datetime = datetime(
+            project.state_changed_at.year, project.state_changed_at.month, 1
+        )
         index: int = self._index[to_count_date]
-        self.categories[project.id][index] += 1 if project.state == "successful" else 0
-
+        self.categories[project.category.id][index] += (
+            1 if project.state == "successful" else 0
+        )
 
 
 @subscribe

diff --git a/kickstarter/processing/monthly_categories_total.py b/kickstarter/processing/monthly_categories_total.py
@@ -1,6 +1,6 @@
 from collections import defaultdict
 from datetime import datetime
-from typing import Dict, List, Optional, Tuple
+from typing import Dict, Optional
 
 from ..models import CategoryModel, ProjectModel
 from . import subscribe
@@ -11,20 +11,23 @@ def __init__(self, start_year=2009, end_year=2020):
         counter: int = 0
         self.dates: Dict[int, datetime] = dict()
         self._index: Dict[datetime, int] = dict()
-        self.categories: Dict[int, Dict[int, int]] = defaultdict(lambda: defaultdict(lambda: 0))
+        self.categories: Dict[int, Dict[int, int]] = defaultdict(
+            lambda: defaultdict(lambda: 0)
+        )
 
-        for year in range(start_year, end_year+1):
+        for year in range(start_year, end_year + 1):
             for month in range(1, 13):
                 date = datetime(year, month, 1)
                 self.dates[counter] = date
                 self._index[date] = counter
                 counter += 1
 
     def count(self, project: ProjectModel):
-        to_count_date: datetime = datetime(project.state_changed_at.year, project.state_changed_at.month, 1)
+        to_count_date: datetime = datetime(
+            project.state_changed_at.year, project.state_changed_at.month, 1
+        )
         index: int = self._index[to_count_date]
-        self.categories[project.id][index] += 1
-
+        self.categories[project.category.id][index] += 1
 
 
 @subscribe

diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -18,6 +18,7 @@ typer = "^0.3.2"
 streamlit = "^0.72.0"
 pydantic = "^1.7.3"
 plotly = "^4.13.0"
+scipy = "^1.5.4"
 
 [tool.poetry.dev-dependencies]
 pytest = "^6.1.2"

diff --git a/requirements.txt b/requirements.txt
@@ -66,6 +66,7 @@ pywinpty==0.5.7; os_name == "nt" and python_version >= "3.6"
 pyzmq==20.0.0; python_version >= "3.6"
 requests==2.25.0; python_version >= "3.6" and python_full_version < "3.0.0" or python_full_version >= "3.5.0" and python_version >= "3.6"
 retrying==1.3.3
+scipy==1.5.4; python_version >= "3.6"
 send2trash==1.5.0; python_version >= "3.6"
 six==1.15.0; python_version >= "3.6" and python_full_version < "3.0.0" or python_full_version >= "3.5.0" and python_version >= "3.6"
 smmap==3.0.4; python_version >= "3.6" and python_full_version < "3.0.0" or python_full_version >= "3.4.0" and python_version >= "3.6"