sketch绘制数据可视化
Analysis of vehicle collision cases of Seattle.
西雅图的车辆碰撞案例分析。
The two terms, data analysis & data visualization seem to have become synonymous, but they are not. A good data expert strives to reach to the analysis through the visualizations. Simple plotting of beautiful graphs covers just the number crunching part, what elevates those visuals is the information and inferences one draws out utilizing statistics and analysis. A large number of people are focusing just on data visualization skills while not necessarily specifying the importance of analytical skills.
数据分析和数据可视化这两个术语似乎已经成为同义词，但事实并非如此。 优秀的数据专家致力于通过可视化工具进行分析。 简单绘制漂亮的图形仅涵盖数字运算部分，而那些可视化则是利用统计和分析得出的信息和推论，从而提升了这些视觉效果。 许多人只专注于数据可视化技能，而不必说明分析技能的重要性。
Data Analysis involves attaching meaning to the data and drawing insights out of it. Data visualization is the process of putting data into a chart, graph, or other visual formats that helps inform analysis and interpretation and not just pretty visuals to look at. Data visuals present the analyzed data in ways that are accessible to and engage different stakeholders done using systematic methods to look for trends, groupings, or other relationships between different types of data.
数据分析涉及将含义附加到数据上并从中汲取见解。 数据可视化是将数据放入图表，图形或其他视觉格式的过程，该过程可帮助进行分析和解释，而不仅仅是看得见的视觉效果。 数据可视化呈现了分析后的数据，使用系统化方法可以找到不同的利益相关者，并使其参与其中，以寻找趋势，分组或不同类型数据之间的其他关系。
The greatest value of a picture is when it forces us to notice what we never expected to see.
图片的最大价值在于它迫使我们注意到我们从未期望看到的事物。
Here we take the example of live accident data to showcase how visualizations integrated with structured inspections to bring out new information and understanding of the scenario at hand. We will dive deep to grasp why certain results have come about and identify opportunities for improvements with further investigation of the data, exploring it as an analyst, beyond producing dashboards.
在这里，我们以现场事故数据为例，展示了可视化如何与结构化检查集成在一起，以带来新的信息和对当前情况的理解。 我们将深入了解为什么会产生某些结果，并通过进一步调查数据，寻找分析员的方式(除了生产仪表板之外)来探索改进的机会。
Our aim here is to understand the causes of accidents, find any patterns persisting in severe accidents, and identify areas of improvement to reduce or prevent such cases. Now we know that more than 38000 people die every year in crashes on U.S. roadways, with a fatality rate of 12.4 deaths per 100 thousand. It is evident that with the increasing number of vehicles in urban and suburban roads, the cases of vehicle accidents are also increasing. We narrowed down our options to the United States’ Emerald City, Seattle. It is a seaport city on the west coast of the US and is the largest city in both the state of Washington and the Pacific Northwest with a living population of ~4 million. We gathered our data from the Seattle Department of Transportation (SDOT). The open-sourced data is regularly updating since 2004. Without further ado, let us jump to analysis.
我们的目的是了解事故原因，发现严重事故中持续存在的任何模式，并确定可以减少或预防此类情况的改进领域。 现在我们知道，每年有38000多人死于美国道路车祸，每10万人中有12.4例死亡。 显然，随着城市和郊区道路上车辆数量的增加，车辆事故的案例也在增加。 我们将选择范围缩小到了美国西雅图的翡翠城。 它是美国西海岸的海港城市，是华盛顿州和西北太平洋地区最大的城市，人口约400万。 我们从西雅图交通局(SDOT)收集了数据。 自2004年以来，开源数据一直在定期更新。事不宜迟，让我们跳到分析。
Photo by Timothy Eberly on Unsplash Timothy Eberly在 Unsplash上 拍摄的照片Although there is no clear trend visible, it seems that cases are declining yearly. There is no pattern of accidents with the months.
尽管没有明显的趋势可见，但案件似乎每年都在下降。 几个月没有发生意外的情况。
To analyze the data based on timestamp, it was helpful to divide the data into 6 equal time buckets.
要基于时间戳分析数据，将数据分为6个相等的时间段很有帮助。
We can see that almost 50% of cases happen between 12:00 P.M. and 08:00 P.M.
我们可以看到几乎50％的案件发生在12:00 PM和08:00 PM之间
But what is the look on collision severity? Although there is no clear correlation between the severity of accidents and time of the day, during late-night, severe accidents are rare. 3 of 4 accidents happening during late-night are just doing property damage.
但是，碰撞严重程度如何？ 尽管事故的严重程度与一天中的时间之间没有明显的相关性，但在深夜，严重的事故很少发生。 深夜发生的4起事故中有3起只是在造成财产损失。
Although just 4% of all the collision cases are because of speeding it is worth noting that the percentage of Injury collisions increases from 26.52% out of the non-speeding collisions to 36.3% in the case of speeding collisions. Similarly, Serious injury collisions increase from 1.39% to 3.92% Likewise, Fatal injury collisions increase from 0.15% to 0.89% because of speeding. While the percentage of Property Damage Only Collisions decreases from 62.14% to 58.87%It shows that indeed, speeding kills more.
尽管只有4％的碰撞案例是由于超速引起的，但值得注意的是，伤害碰撞的百分比从非超速碰撞中的26.52％增加到超速碰撞中的36.3％。 同样，严重伤害碰撞从1.39％增加到3.92％同样，致命伤害碰撞由于超速而从0.15％增加到0.89％。 仅有碰撞的财产损失百分比从62.14％降低到58.87％，这表明，超速驾驶确实可以杀死更多人。
Speeding cases distributed with time shows that there are a similar number of cases in Evening, Late Night, and Noon. But percentage-wise distribution tells some different story. 11.05% of all accidents happening in late-night are because of speeding, indicating speeding accidents are more likely to occur at night due to less visibility.
随时间分布的超速案件表明，晚上，深夜和中午的案件数量相似。 但是按百分比分配的情况却有所不同。 在深夜发生的所有事故中，有11.05％是由超速引起的，这表明由于可见度较低，超速驾驶事故更有可能在晚上发生。
While analyzing collision type, we noticed that majority of speeding collisions are due to motor vehicles striking another motor vehicle from the rear end (which was expected).
在分析碰撞类型时，我们注意到大多数超速碰撞是由于机动车辆从后端撞击另一辆机动车辆(这是预期的)。
In the search of speeding collision hotspots, we found “Battery Street Tunnel” to be on the top of the list.
在寻找超速碰撞热点时，我们发现“ Battery Street Tunnel”位于列表的顶部。
https://bridgehunter.com/photos/19/72/197274-M.jpg https://bridgehunter.com/photos/19/72/197274-M.jpgDistraction and inattention pose a considerable threat to road safety, not only for car drivers but also for vulnerable road users. Our data shows that ~18% of collisions happening in the morning and early morning are due to inattention of drivers.
分心和注意力不集中对道路安全构成了很大的威胁，不仅对汽车驾驶员，而且对脆弱的道路使用者也构成威胁。 我们的数据显示，在早晨和清晨发生的约18％的碰撞是由于驾驶员疏忽引起的。
We noticed that 35% of collisions happening because of inattention are rear-ended collisions.
我们注意到，由于注意力不集中而发生的35％的碰撞是追尾碰撞。
Alcohol/drug-influenced accidents are becoming a huge problem for metropolitans. It is quite evident from the data of total alcoholic and drug-induced collisions, ~64% of cases happen during Night and Late Night.
对于大都会人来说，酒精/毒品影响的事故正成为一个巨大的问题。 从酒精和药物引起的全部碰撞的数据中可以明显看出，约有64％的病例发生在夜间和深夜。
We can see there are relatively more cases of alcoholic and drug-influenced collision during December (Probably because of the holiday season)
我们可以看到12月期间发生的酒精和药物影响的撞车事件相对较多(可能是由于假期)
ALASKAN VI SB BETWEEN COLUMBIA ST ON RP AND ALASKAN WY VI SB EFR OFF RP has the greatest number of cases of under influence collisions
RP上的哥伦比亚ST和阿拉斯加的WY VI SB EFR OFF之间的ALASKAN VI SB受撞击影响的案例数量最多
We analyzed the distribution of collisions by the collision type. Collision type is further bifurcated by the SDOT collision code. It is a unique code that describes how the collision occurred.
我们根据碰撞类型分析了碰撞的分布。 SDOT冲突代码进一步将冲突类型分为两类。 它是描述冲突如何发生的唯一代码。
SDOT Collision Code Snapshot SDOT冲突代码快照People are more likely to hit Parked cars during Night. Motor vehicles tend to hit fixed objects more during late night-time, again a cause of low visibility.
人们在夜间更有可能撞到停放的汽车。 机动车在深夜往往会更多地撞击固定物体，这也是能见度低的原因。
During this analysis, we came across a striking insight. There is ~65% more chance to encounter a collision while taking a Left turn than taking a Right turn. This stat indicates that people are more complacent while taking a left turn than taking a right turn. This might be because in a right turn mostly they cross a lane at the intersection to reach the other side, thus requiring more attention.
在分析过程中，我们遇到了惊人的见解。 与左转弯相比，左转弯碰到碰撞的机率高出65％。 该统计数据表明，与左转弯相比，人们左转弯时更自满。 这可能是因为大多数情况下，他们在右转弯时穿过交叉路口的车道到达另一侧，因此需要更多注意。
We can see ~40% of collisions are of a motor vehicle striking another motor vehicle’s front end at an angle.
我们可以看到，约40％的碰撞是汽车以一定角度撞击另一辆汽车的前端。
Collision at Blocks and Intersections are more common than in Alley.
街区和路口的碰撞比胡同中的碰撞更为普遍。
While analyzing the accidents, we came across certain accident hotspots. “6th AVE AND JAMES ST” has an abnormally high number of collisions since 2016
在分析事故时，我们遇到了某些事故热点。 自2016年以来，“ 6th AVE AND JAMES ST”发生了异常多的碰撞
We have data on-road and light conditions where the collision happened. We can see that majority of collisions happened at dry roads and in daylight. We checked for the severity of these conditions.
我们拥有发生碰撞的道路和光照条件的数据。 我们可以看到，大多数碰撞发生在干燥的道路和白天。 我们检查了这些情况的严重性。
In 2020, the whole world faced a pandemic of the COVID-19 outbreak. Its effect on travel behavior has also been sudden and wide-ranging. We can see that there is a sudden drop in accident cases in 2020, due to the COVID-19 pandemic
2020年，全世界都面临着COVID-19爆发的大流行。 它对旅行行为的影响也是突然而广泛的。 我们可以看到，由于COVID-19大流行，到2020年事故发生率突然下降
After the thorough analysis, we have a better understanding of the causes and reasons for major road accidents. We see that although only 26% of accidents involve any injury, the number is still too high. The majority of the cases happen between the afternoon and evening, thus people driving during these hours should be more careful. The visuals convey to us that severity of the accident is higher during the day than late night, something we thought wasn’t likely. We also see that speeding causes more accidents late at night. Additionally, in night-time accidents primarily happen because of less visibility. Thus, drivers should avoid speeding and be more careful during the night to avoid accidents. The data tells us that severe accidents more likely happen at blocks or intersections, as expected.
经过深入分析，我们对重大道路交通事故的原因和原因有了更好的了解。 我们看到，尽管只有26％的事故涉及任何伤害，但这一数字仍然很高。 大多数情况发生在下午和晚上之间，因此在这些时间驾驶的人们应格外小心。 视觉效果向我们传达出事故的严重程度在白天比深夜更高，这是我们认为不可能的。 我们还看到，超速驾驶会在深夜造成更多事故。 另外，在夜间发生事故主要是因为能见度较低。 因此，驾驶员应避免超速驾驶，并在夜间多加注意避免发生事故。 数据告诉我们，按预期，重大事故更有可能在街区或交叉路口发生。
Moreover, we see that there is a spike in drunken driving cases during December (holiday season), the law enforcement should particularly take care during this time for such incidents. Here we see that utilizing analysis on top of the visuals entails one to bring out insights that may lead to a breakthrough, for example understanding that the chances of being in an accident while turning left are more than while turning right. This information forms the core of the data analysis process. An effective, well-designed, and beautiful visualization is great but you shouldn’t lose the context hidden in data art and must focus upon data-driven decisions that may engage your stakeholders. It is crucial to be a good visual storyteller rather than a visual artist, bringing out stories from data is what companies seek and require.
此外，我们发现在12月(节假日)酒后驾驶案件激增，执法部门在此期间应特别注意此类事件。 在这里，我们看到，利用视觉效果之上的分析需要带来洞察力，这可能会带来突破，例如，了解到左转弯时发生事故的机率要大于右转弯时的机率。 这些信息构成了数据分析过程的核心。 有效，精心设计且美观的可视化效果很好，但您不应失去隐藏在数据艺术中的上下文，而必须专注于数据驱动的决策，这些决策可能会使您的利益相关者参与其中。 成为一名优秀的视觉讲故事者而不是视觉艺术家至关重要，从数据中找出故事是公司寻求和要求的。
Find the complete analysis at:https://www.kaggle.com/anmoltripathi/vehicle-collision-data-analysis-sdot
在以下位置找到完整的分析： https : //www.kaggle.com/anmoltripathi/vehicle-collision-data-analysis-sdot
Need such data-driven solutions for your business?
需要为您的企业提供此类数据驱动的解决方案吗？
Visit us at www.intellyticssolutions.com
请访问我们的 网站www.intellyticssolutions.com
翻译自: https://medium.com/intellyticssolutions/data-analysis-beyond-plotting-of-colourful-visualisations-35886facad21
sketch绘制数据可视化
相关资源：四史答题软件安装包exe