1.采集中国地震局地震数据约100万条存入.csv和mysql,清洗后的.csv上传mysql;
3.分析指标离线可选用Hive,实时可选装PySpark/PyFlink,可三选一也可以只选一种或者三个都选；
4.计算结果使用sqoop工具对接到mysql数据库的指标表；
5.使用flask+echarts制作可视化大屏、layui查询表格；
6.使用卷积神经网络KNN CNN RNN对地震数据进行预测；
创新点：全新DrssionPage爬虫框架、可视化大屏、离线计算实时计算全部实现、深度学习算法预测地震。
核心算法代码分享如下：
import sys
from db import cnn
import pandas as pd
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegressiondef predict1():sql = "select `stime_year`, `num` as value from table02" \" order by stime_year asc "with cnn.cursor() as cursor:cursor.execute(sql)print(sql)names = []y = []for line in cursor.fetchall():# print(line)y.append(line[1])names.append(line[0])y = y[::-1]X = [1, 2, 3, 4, 5]X = pd.DataFrame(X)X = X.valuesPoly_regressor = PolynomialFeatures(degree=2)Poly_X = Poly_regressor.fit_transform(X)regressor = LinearRegression()regressor.fit(Poly_X, y)p1 = regressor.predict(Poly_regressor.fit_transform([[8]]))p2 = regressor.predict(Poly_regressor.fit_transform([[9]]))p3 = regressor.predict(Poly_regressor.fit_transform([[10]]))r = []r.append(round(float(p1[0]),2))r.append(round(float(p2[0]),2))r.append(round(float(p3[0]),2))return rdef predict2():sql = "select `stime_year`, `num` as value from table02" \" order by stime_year asc "with cnn.cursor() as cursor:cursor.execute(sql)print(sql)names = []y = []for line in cursor.fetchall():# print(line)y.append(line[1])names.append(line[0])y = y[::-1]X = [1, 2, 3, 4, 5]X = pd.DataFrame(X)X = X.valuesPoly_regressor = PolynomialFeatures(degree=2)Poly_X = Poly_regressor.fit_transform(X)regressor = LinearRegression()regressor.fit(Poly_X, y)p1 = regressor.predict(Poly_regressor.fit_transform([[8]]))p2 = regressor.predict(Poly_regressor.fit_transform([[9]]))p3 = regressor.predict(Poly_regressor.fit_transform([[10]]))r = []r.append(round(float(p1[0]),2))r.append(round(float(p2[0]),2))r.append(round(float(p3[0]),2))return rdef predict3():sql = "select `stime_year`, `num` as value from table02" \" order by stime_year asc "with cnn.cursor() as cursor:cursor.execute(sql)print(sql)names = []y = []for line in cursor.fetchall():# print(line)y.append(line[1])names.append(line[0])y = y[::-1]X = [1, 2, 3, 4, 5]X = pd.DataFrame(X)X = X.valuesPoly_regressor = PolynomialFeatures(degree=2)Poly_X = Poly_regressor.fit_transform(X)regressor = LinearRegression()regressor.fit(Poly_X, y)p1 = regressor.predict(Poly_regressor.fit_transform([[8]]))p2 = regressor.predict(Poly_regressor.fit_transform([[9]]))p3 = regressor.predict(Poly_regressor.fit_transform([[10]]))r = []r.append(round(float(p1[0]),2))r.append(round(float(p2[0]),2))r.append(round(float(p3[0]),2))return rdef predict4():sql = "select `stime_year`, `num` as value from table02" \" order by stime_year asc "with cnn.cursor() as cursor:cursor.execute(sql)print(sql)names = []y = []for line in cursor.fetchall():# print(line)y.append(line[1])names.append(line[0])y = y[::-1]X = [1, 2, 3, 4, 5]X = pd.DataFrame(X)X = X.valuesPoly_regressor = PolynomialFeatures(degree=2)Poly_X = Poly_regressor.fit_transform(X)regressor = LinearRegression()regressor.fit(Poly_X, y)p1 = regressor.predict(Poly_regressor.fit_transform([[8]]))p2 = regressor.predict(Poly_regressor.fit_transform([[9]]))p3 = regressor.predict(Poly_regressor.fit_transform([[10]]))r = []r.append(round(float(p1[0]),2))r.append(round(float(p2[0]),2))r.append(round(float(p3[0]),2))return rif __name__ == '__main__':#name = sys.argv[1]ret = []r1 = predict1()r2 = predict2()r3 = predict3()r4 = predict4()ret.append(r1)ret.append(r2)ret.append(r3)ret.append(r4)print(ret)print(r1)print(abs(int(r1[0])))print(abs(int(r1[1])))print(abs(int(r1[2])))sql_day01="replace into table02(stime_year,num) values (%s,%s)"data_day01 =('2024(预测)',abs(int(r1[0])))sql_day02="replace into table02(stime_year,num) values (%s,%s)"data_day02 = ('2025(预测)', abs(int(r1[1])))sql_day03="replace into table02(stime_year,num) values (%s,%s)"data_day03 = ('2026(预测)', abs(int(r1[2])))cur=cnn.cursor()cur.execute(sql_day01,data_day01)cur.execute(sql_day02,data_day02)cur.execute(sql_day03,data_day03)cnn.commit()cur.close()
)
